Camera-based 3D Semantic Occupancy Prediction (SOP) is crucial for understanding complex 3D scenes from limited 2D image observations. Existing SOP methods typically aggregate contextual features to assist the occupancy representation learning, alleviating issues like occlusion or ambiguity. However, these solutions often face misalignment issues wherein the corresponding features at the same position across different frames may have different semantic meanings during the aggregation process, which leads to unreliable contextual fusion results and an unstable representation learning process. To address this problem, we introduce a new Hierarchical context alignment paradigm for a more accurate SOP (Hi-SOP). Hi-SOP first disentangles the geometric and temporal context for separate alignment, which two branches are then composed to enhance the reliability of SOP. This parsing of the visual input into a local-global alignment hierarchy includes: (I) disentangled geometric and temporal separate alignment, within each leverages depth confidence and camera pose as prior for relevant feature matching respectively; (II) global alignment and composition of the transformed geometric and temporal volumes based on semantics consistency. Our method outperforms SOTAs for semantic scene completion on the SemanticKITTI & NuScenes-Occupancy datasets and LiDAR semantic segmentation on the NuScenes dataset.

We propose Radar-Camera fusion transformer (RaC-Former) to boost the accuracy of 3D object detection by the following insight. The Radar-Camera fusion in outdoor 3D scene perception is capped by the image-to-BEV transformation-if the depth of pixels is not accurately estimated, the naive combination of BEV features actually integrates unaligned visual content. To avoid this problem, we propose a query-based framework that enables adaptive sampling of instance-relevant features from both the bird’s-eye view (BEV) and the original image view. Furthermore, we enhance system performance by two key designs: optimizing query initialization and strengthening the representational capacity of BEV. For the former, we introduce an adaptive circular distribution in polar coordinates to refine the initialization of object queries, allowing for a distance-based adjustment of query density. For the latter, we initially incorporate a radar-guided depth head to refine the transformation from image view to BEV. Subsequently, we focus on leveraging the Doppler effect of radar and introduce an implicit dynamic catcher to capture the temporal elements within the BEV. Extensive experiments on nuScenes and View-of-Delft (VoD) datasets validate the merits of our design. Remarkably, our method achieves superior results of 64.9% mAP and 70.2% NDS on nuScenes. RaCFormer also secures the state-of-the-art performance on the VoD dataset. Code is available at https://github.com/cxmomo/RaCFormer.

Current 3D Large Multimodal Models (3D LMMs) have shown tremendous potential in 3D-vision-based dialogue and reasoning. However, how to further enhance 3D LMMs to achieve fine-grained scene understanding and facilitate flexible human-agent interaction remains a challenging problem. In this work, we introduce 3D-LLaVA, a simple yet highly powerful 3D LMM designed to act as an intelligent assistant in comprehending, reasoning, and interacting with the 3D world. Unlike existing top-performing methods that rely on complicated pipelines—such as offline multi-view feature extraction or additional task-specific heads—3D-LLaVA adopts a minimalist design with integrated architecture and only takes point clouds as input. At the core of 3D-LLaVA is a new Omni Superpoint Transformer (OST), which integrates three functionalities: (1) a visual feature selector that converts and selects visual tokens, (2) a visual prompt encoder that embeds interactive visual prompts into the visual token space, and (3) a referring mask decoder that produces 3D masks based on text description. This versatile OST is empowered by the hybrid pretraining to obtain perception priors and leveraged as the visual connector that bridges the 3D data to the LLM. After performing unified instruction tuning, our 3D-LLaVA reports impressive results on various benchmarks. The code and model will be released at https://github.com/djiajunustc/3D-LLaVA.

Discrete tokens provide compact and domain-adaptable representations of speech features. However, their application to disordered speech, characterized by articulation imprecision and significant mismatch with normal voice, remains unexplored. To this end, this paper proposes novel phone-purity guided (PPG) discrete tokens to address the weakened phonetic discrimination arising during unsupervised K-means clustering or vector quantization of continuous features. Phonetic label supervision is incorporated to regularize the maximum likelihood and reconstruction error costs in standard K-means and VAE-VQ-based token extraction. Experiments on the UASpeech corpus show that PPG-based discrete tokens extracted from HuBERT consistently outperform hybrid TDNN and End-to-End (E2E) Conformer systems using non-PPG tokens. Statistically significant word error rate (WER) reductions of up to 0.99% and 1.77% absolute (3.21% and 4.82% relative) are achieved across varying codebook sizes for the 16 UASpeech test dysarthric speakers. The lowest WER of 23.25% is obtained by combining systems using complementary token features. Consistent improvements are also observed in phone purity, and t-SNE visualizations demonstrate sharper decision boundaries between K-means/VAE-VQ clusters with the introduction of phone-purity guidance.

This paper presents a novel mixed-precision quantization approach for speech foundation models that tightly integrates mixed-precision learning and quantized model parameter estimation into one single model compression stage. Experiments conducted on LibriSpeech dataset with fine-tuned wav2vec2.0-base and HuBERT-large models suggest the resulting mixed-precision quantized models increased the lossless compression ratio by factors up to 1.7x and 1.9x over the respective uniform-precision and two-stage mixed-precision quantized baselines that perform precision learning and model parameters quantization in separate and disjointed stages, while incurring no statistically word error rate (WER) increase over the 32-bit full-precision models. The system compression time of wav2vec2.0-base and HuBERT-large models is reduced by up to 1.9 and 1.5 times over the two-stage mixed-precision baselines, while both produce lower WERs. The best-performing 3.5-bit mixed-precision quantized HuBERT-large model produces a lossless compression ratio of 8.6x over the 32-bit full-precision system.

Flexible instruction-guided 6-DoF grasping is a significant yet challenging task for real-world robotic systems. Existing methods utilize the contextual understanding capabilities of the large language models (LLMs) to establish mappings between expressions and targets, allowing robots to comprehend users' intentions in the instructions. However, the LLM's knowledge about objects' physical properties remains underexplored despite its tight relevance to grasping. In this work, we propose GraspCoT, a 6-DoF grasp detection framework that integrates a Chain-of-Thought (CoT) reasoning mechanism oriented to physical properties, guided by auxiliary question-answering (QA) tasks. Particularly, we design a set of QA templates to enable hierarchical reasoning that includes three stages: target parsing, physical property analysis, and grasp action selection. Moreover, GraspCoT presents a unified multimodal LLM architecture, which encodes multi-view observations of 3D scenes into 3D-aware visual tokens, and then jointly embeds these visual tokens with CoT-derived textual tokens within LLMs to generate grasp pose predictions. Furthermore, we present IntentGrasp, a large-scale benchmark that fills the gap in public datasets for multi-object grasp detection under diverse and indirect verbal commands. Extensive experiments on IntentGrasp demonstrate the superiority of our method, with additional validation in real-world robotic applications confirming its practicality. The code is available at https://github.com/cxmomo/GraspCoT.

A major breakthrough in 3D reconstruction is the feedforward paradigm to generate pixel-wise 3D points or Gaussian primitives from sparse, unposed images. To further incorporate semantics while avoiding the significant memory and storage costs of high-dimensional semantic features, existing methods extend this paradigm by associating each primitive with a compressed semantic feature vector. However, these methods have two major limitations: (a) the naively compressed feature compromises expressiveness, affecting the model's ability to capture fine-grained semantics, and (b) the pixel-wise primitive prediction introduces redundancy in overlapping areas, causing unnecessary memory overhead. To this end, we introduce \textbf{SpatialSplat}, a feedforward framework that produces redundancy-aware Gaussians and capitalizes on a dual-field semantic representation. Particularly, with the insight that primitives within the same instance exhibit high semantic consistency, we decompose the semantic representation into a coarse feature field that encodes uncompressed semantics with minimal primitives, and a fine-grained yet low-dimensional feature field that captures detailed inter-instance relationships. Moreover, we propose a selective Gaussian mechanism, which retains only essential Gaussians in the scene, effectively eliminating redundant primitives. Our proposed Spatialsplat learns accurate semantic information and detailed instances prior with more compact 3D Gaussians, making semantic 3D reconstruction more applicable. We conduct extensive experiments to evaluate our method, demonstrating a remarkable 60\% reduction in scene representation parameters while achieving superior performance over state-of-the-art methods. The code is available at https://github.com/shengyuuu/SpatialSplat.git

Recently, 3D Gaussian Splatting (3DGS) has reshaped the field of photorealistic 3D reconstruction, achieving impressive rendering quality and speed. However, when applied to large-scale street scenes, existing methods suffer from rapidly escalating per-viewpoint reconstruction costs as scene size increases, leading to significant computational overhead. After revisiting the conventional pipeline, we identify three key factors accounting for this issue: unnecessary local-to-global transformations, excessive 3D-to-2D projections, and inefficient rendering of distant content. To address these challenges, we propose S3R-GS, a 3DGS framework that Streamlines the pipeline for large-scale Street Scene Reconstruction, effectively mitigating these limitations. Moreover, most existing street 3DGS methods rely on ground-truth 3D bounding boxes to separate dynamic and static components, but 3D bounding boxes are difficult to obtain, limiting real-world applicability. To address this, we propose an alternative solution with 2D boxes, which are easier to annotate or can be predicted by off-the-shelf vision foundation models. Such designs together make S3R-GS readily adapt to large, in-the-wild scenarios. Extensive experiments demonstrate that S3R-GS enhances rendering quality and significantly accelerates reconstruction. Remarkably, when applied to videos from the challenging Argoverse2 dataset, it achieves state-of-the-art PSNR and SSIM, reducing reconstruction time to below 50%--and even 20%--of competing methods.

Music similarity retrieval is fundamental for managing and exploring relevant content from large collections in streaming platforms. This paper presents a novel cross-modal contrastive learning framework that leverages the open-ended nature of text descriptions to guide music similarity modeling, addressing the limitations of traditional uni-modal approaches in capturing complex musical relationships. To overcome the scarcity of high-quality text-music paired data, this paper introduces a dual-source data acquisition approach combining online scraping and LLM-based prompting, where carefully designed prompts leverage LLMs’ comprehensive music knowledge to generate contextually rich descriptions. Extensive experiments demonstrate that the proposed framework achieves significant performance improvements over existing benchmarks through objective metrics, subjective evaluations, and real-world A/B testing on the Huawei Music streaming platform.

Autoregressive models have emerged as a powerful generative paradigm for visual generation. The current de-facto standard of next token prediction commonly operates over a single-scale sequence of dense image tokens, and is incapable of utilizing global context especially for early tokens prediction. In this paper, we introduce a new autoregressive design to model a hierarchy from a few low-resolution image tokens to the typical dense image tokens, and delve into a thorough hierarchical dependency across multi-scale image tokens. Technically, we present a Hierarchical Masked Autoregressive models (Hi-MAR) that pivot on low-resolution image tokens to trigger hierarchical autoregressive modeling in a multi-phase manner. Hi-MAR learns to predict a few image tokens in low resolution, functioning as intermediary pivots to reflect global structure, in the first phase. Such pivots act as the additional guidance to strengthen the next autoregressive modeling phase by shaping global structural awareness of typical dense image tokens. A new Diffusion Transformer head is further devised to amplify the global context among all tokens for mask token prediction. Extensive evaluations on both class-conditional and text-to-image generation tasks demonstrate that Hi-MAR outperforms typical AR baselines, while requiring fewer computational costs.

In recent years, weakly supervised object detection (WSOD) has attracted much attention due to its low labeling cost. The success of recent WSOD models is often ascribed to the two-stage multi-class classification (MCC) task, i.e., multiple instance learning and online classification refinement. Despite achieving non-trivial progresses, these methods overlook potential classification ambiguities between these two MCC tasks and fail to leverage their unique strengths. In this work, we introduce a novel WSOD framework to ameliorate these two issues. For one thing, we propose a self-classification enhancement module that integrates intra-class binary classification (ICBC) to bridge the gap between the two distinct MCC tasks. The ICBC task enhances the network’s discrimination between positive and mis-located samples in a class-wise manner and forges a mutually reinforcing relationship with the MCC task. For another, we propose a self-classification correction algorithm during inference, which combines the results of both MCC tasks to effectively reduce the mis-classified predictions. Extensive experiments on the prevalent VOC 2007 & 2012 datasets demonstrate the superior performance of our framework.

Integrating large language models (LLMs) into autonomous driving motion planning has recently emerged as a promising direction, offering enhanced interpretability, better controllability, and improved generalization in rare and long-tail scenarios. However, existing methods often rely on abstracted perception or map-based inputs, missing crucial visual context, such as fine-grained road cues, accident aftermath, or unexpected obstacles, which are essential for robust decision-making in complex driving environments. To bridge this gap, we propose VLMPlanner, a hybrid framework that combines a learning-based real-time planner with a vision-language model (VLM) capable of reasoning over raw images. The VLM processes multi-view images to capture rich, detailed visual information and leverages its common-sense reasoning capabilities to guide the real-time planner in generating robust and safe trajectories. Furthermore, we develop the Context-Adaptive Inference Gate (CAI-Gate) mechanism that enables the VLM to mimic human driving behavior by dynamically adjusting its inference frequency based on scene complexity, thereby achieving an optimal balance between planning performance and computational efficiency. We evaluate our approach on the large-scale, challenging nuPlan benchmark, with comprehensive experimental results demonstrating superior planning performance in scenarios with intricate road conditions and dynamic elements.

In this work, we present DeepEraser, an effective deep network for generic text removal. DeepEraser utilizes a recurrent architecture that erases the text in an image via iterative operations. Our idea comes from the process of erasing pencil script, where the text area designated for removal is subject to continuous monitoring and the text is attenuated progressively, ensuring a thorough and clean erasure. Technically, at each iteration, an innovative erasing module is deployed, which not only explicitly aggregates the previous erasing progress but also mines additional semantic context to erase the target text. Through iterative refinements, the text regions are progressively replaced with more appropriate content and finally converge to a relatively accurate status. Furthermore, a custom mask generation strategy is introduced to improve the capability of DeepEraser for adaptive text removal, as opposed to indiscriminately removing all the text in an image. Our DeepEraser is notably compact with only 1.4 M parameters and trained in an end-to-end manner. To verify its effectiveness, extensive experiments are conducted on several prevalent benchmarks, including SCUT-Syn, SCUT-EnsText, and Oxford Synthetic text dataset. The quantitative and qualitative results demonstrate the effectiveness of our DeepEraser over the state-of-the-art methods, as well as its strong generalization ability in custom mask text removal.

We present that visual grounding and image captioning, which perform as two mutually inverse processes, can be bridged together for collaborative training by careful designs. By consolidating this idea, we introduce CyCo, a cyclic-consistent learning framework to ameliorate the independent training pipelines of visual grounding and image captioning. The proposed framework (1) allows the semi-weakly supervised training of visual grounding; (2) improves the performance of fully supervised visual grounding; (3) yields a general captioning model that can describe arbitrary image regions. Extensive experiments show that our fully supervised grounding model achieves state-of-the-art performance, and the semi-weakly supervised one also exhibits competitive performance compared to the fully supervised counterparts. Our image captioning model has the capability to freely describe image regions and meanwhile shows impressive performance on prevalent captioning benchmarks.

In this paper, we focus on the open-set panoptic segmentation (OPS) task to circumvent the data explosion problem. Different from the close-set setting, OPS targets to detect both known and unknown categories, where the latter is not annotated during training. Different from existing work that only selects a few common categories as unknown ones, we move forward to the real-world scenario by considering the various tail categories (~1k). To this end, we first build a new dataset with long-tail distribution for the OPS task. Based on this dataset, we additionally add a new class type for unknown classes and re-define the training annotations to make the OPS definition more complete and reasonable. Moreover, we analyze the influence of several significant factors in the OPS task and explore the upper bound of performance on unknown classes with different settings. Furthermore, based on the analyses, we design an effective two-phase framework for the OPS task, including thing-agnostic map generation and unknown segment mining. We further adopt semi-supervised learning to improve the OPS performance. Experimental results on different datasets validate the effectiveness of our method.

Practical application of model-based speaker adaptation techniques to end-to-end ASR systems is hindered by speaker-level data scarcity and latency in speaker-dependent (SD) parameters update. To this end, data-efficient and low-latency rapid feature-based speaker adaptation approaches are proposed in this paper for state-of-the-art Conformer ASR systems. Compact subspace projection of training data estimated SD hidden layer output scaling or bias parameters is used to represent the most distinctive speaker "bases". A feature-driven prediction network containing purpose-built speaker-aware memory is designed to on-the-fly produce homogeneous SD basis interpolation, and facilitate rapid speaker adaptation. Experimental results on the 300-hr Switchboard corpus suggest that the proposed adaptation approach produces statistically significant word error rate (WER) reductions of up to 1.0% absolute (8.4% relative) over the baseline speaker-independent and i-vector adapted Conformers before and after external LM rescoring. Consistent WER reductions of up to 2.0% absolute (16.3% relative) and real-time factor speeding up ratios of up to 10.9 times are also obtained over offline model-based adaptation across different speaker-level data quantity operating points. T-SNE visualization reveals the on-the-fly predicted SD basis weights present intuitively more consistent speaker features than i-vectors.

Automatic recognition of disordered speech remains a highly challenging task to date due to data scarcity. This paper presents a reinforcement learning (RL) based on-the-fly data augmentation approach for training state-of-the-art PyChain TDNN and end-to-end Conformer ASR systems on such data. The handcrafted temporal and spectral mask operations in the standard SpecAugment method that are task and system dependent, together with additionally introduced minimum and maximum cut-offs of these masks, are now automatically learned using an RNN-based policy controller and tightly integrated with ASR system training. Experiments on the UASpeech corpus suggest the proposed RL-based data augmentation consistently produced performance superior or comparable to that obtained using expert or handcrafted SpecAugment policies. Our RL auto-augmented PyChain TDNN system produced an overall WER of 28.79% on the UASpeech test set of 16 dysarthric speakers.

Interspeech 2024, 1-5 September 2024, Kos, Greece

We propose a novel one-pass multiple ASR systems joint compression and quantization approach using an all-in-one neural model. A single compression cycle allows multiple nested systems with varying Encoder depths, widths, and quantization precision settings to be simultaneously constructed without the need to train and store individual target systems separately. Experiments consistently demonstrate the multiple ASR systems compressed in a single all-in-one model produced a word error rate (WER) comparable to, or lower by up to 1.01% absolute (6.98% relative) than individually trained systems of equal complexity. A 3.4x overall system compression and training time speed-up was achieved. Maximum model size compression ratios of 12.8x and 3.93x were obtained over the baseline Switchboard-300hr Conformer and LibriSpeech-100hr fine-tuned wav2vec2.0 models, respectively, incurring no statistically significant WER increase.

This paper proposes a novel non-autoregressive (NAR) block-based Attention Mask Decoder (AMD) that flexibly balances performance-efficiency trade-offs for Conformer ASR systems. AMD performs parallel NAR inference within contiguous blocks of output labels that are concealed using attention masks, while conducting left-to-right AR prediction and history context amalgamation between blocks. A beam search algorithm is designed to leverage a dynamic fusion of CTC, AR Decoder, and AMD probabilities. Experiments on the LibriSpeech-100hr corpus suggest the tripartite Decoder incorporating the AMD module produces a maximum decoding speed-up ratio of 1.73x over the baseline CTC+AR decoding, while incurring no statistically significant word error rate (WER) increase on the test sets. When operating with the same decoding real time factors, statistically significant WER reductions of up to 0.7% and 0.3% absolute (5.3% and 6.1% relative) were obtained over the CTC+AR baseline.

The recent advances in query-based multi-camera 3D object detection are featured by initializing object queries in the 3D space, and then sampling features from perspective-view images to perform multi-round query refinement. In such a framework, query points near the same camera ray are likely to sample similar features from very close pixels, resulting in ambiguous query features and degraded detection accuracy. To this end, we introduce RayFormer, a camera-ray-inspired query-based 3D object detector that aligns the initialization and feature extraction of object queries with the optical characteristics of cameras. Specifically, RayFormer transforms perspective-view image features into bird's eye view (BEV) via the lift-splat-shoot method and segments the BEV map to sectors based on the camera rays. Object queries are uniformly and sparsely initialized along each camera ray, facilitating the projection of different queries onto different areas in the image to extract distinct features. Besides, we leverage the instance information of images to supplement the uniformly initialized object queries by further involving additional queries along the ray from 2D object detection boxes. To extract unique object-level features that cater to distinct queries, we design a ray sampling method that suitably organizes the distribution of feature sampling points on both images and bird's eye view. Extensive experiments are conducted on the nuScenes dataset to validate our proposed ray-inspired model design. The proposed RayFormer achieves 55.5% mAP and 63.3% NDS, respectively.

Fusing the data of millimeter-wave Radar sensors and high-definition cameras has emerged as a viable approach to achieving precise 3D object detection for roadside traffic surveillance. For roadside perception systems, earlier studies have pointed out that it is better to perform the fusion on the 2D image plane than on the BEV plane (which is popular for on-car perception systems), especially when the perception range is large (e.g., >150m). Image-plane fusion requires critical transformations, like perspective projection from the Radar's BEV to the camera's 2D plane and reverse IPM. However, real-world issues like uneven terrain and sensor movement degrade these transformations' precision, impacting fusion effectiveness. To alleviate these issues, we propose a geometry-based Radar-camera fusion method on the ground, namely FARFusion V2. Specifically, we extend the ground-plane assumption in FARFusion[20] to support arbitrary shapes by formulating the ground height as an implicit representation based on geometric transformations. By incorporating the ground information, we can enhance Radar data with target height measurements. Consequently, we can thus project the enhanced Radar data onto the 2D plane to obtain more accurate depth information, thereby assisting the IPM process. A real-time parameterized transformation parameters estimation module is further introduced to refine the view transformation processes. Moreover, considering various measurement noises across these two sensors, we introduce an uncertainty-based depth fusion strategy into the 2D fusion process to maximize the probability of obtaining the optimal depth value. Extensive experiments are conducted on our collected roadside OWL benchmark, demonstrating the excellent localization capacity of FARFusion V2 in far-range scenarios. Our method achieves an average location accuracy of 0.771m when we extend the detection range up to 500m.

Self-supervised learning (SSL) based speech foundation models have been applied to a wide range of ASR tasks. However, their application to dysarthric and elderly speech via data-intensive parameter fine-tuning is confronted by in-domain data scarcity and mismatch. To this end, this paper explores a series of approaches to integrate domain fine-tuned SSL pre-trained models and their features into TDNN and Conformer ASR systems for dysarthric and elderly speech recognition. These include: a) input feature fusion between standard acoustic frontends and domain fine-tuned SSL speech representations; b) frame-level joint decoding between TDNN systems separately trained using standard acoustic features alone and those with additional domain fine-tuned SSL features; and c) multi-pass decoding involving the TDNN/Conformer system outputs to be rescored using domain fine-tuned pre-trained ASR models. In addition, fine-tuned SSL speech features are used in acoustic-to-articulatory (A2A) inversion to construct multi-modal ASR systems. Experiments are conducted on four tasks: the English UASpeech and TORGO dysarthric speech corpora; and the English DementiaBank Pitt and Cantonese JCCOCC MoCA elderly speech datasets. The TDNN systems constructed by integrating domain-adapted HuBERT, wav2vec2-conformer or multi-lingual XLSR models and their features consistently outperform the standalone fine-tuned SSL pre-trained models. These systems produced statistically significant WER or CER reductions of6.53%,1.90%,2.04%and7.97%absolute (24.10%,23.84%,10.14%and31.39%relative) on the four tasks respectively. Consistent improvements in Alzheimer's Disease detection accuracy are also obtained using the DementiaBank Pitt elderly speech recognition outputs.

Despite the rapid progress of automatic speech recognition (ASR) technologies targeting normal speech, accurate recognition of dysarthric and elderly speech remains a highly challenging task to date. It is difficult to collect large quantities of such data for ASR system development due to the mobility issues often found among these users. To this end, data augmentation techniques play a vital role. In contrast to existing data augmentation techniques only modifying the speaking rate or overall shape of spectral contour, fine-grained spectro-temporal differences between dysarthric, elderly and normal speech are modelled using a novel set of speaker-dependent (SD) generative adversarial networks (GAN) based data augmentation approaches in this article. These flexibly allow both: a) temporal or speed perturbed normal speech spectra to be modified and closer to those of an impaired speaker when parallel speech data is available; and b) for non-parallel data, the SVD decomposed normal speech spectral basis features to be transformed into those of a target elderly speaker before being re-composed with the temporal bases to produce the augmented data for state-of-the-art TDNN and Conformer ASR system training. Experiments are conducted on four tasks: the English UASpeech and TORGO dysarthric speech corpora; the English DementiaBank Pitt and Cantonese JCCOCC MoCA elderly speech datasets. The proposed GAN based data augmentation approaches consistently outperform the baseline speed perturbation method by up to 0.91% and 3.0% absolute (9.61% and 6.4% relative) WER reduction on the TORGO and DementiaBank data respectively. Consistent performance improvements are retained after applying LHUC based speaker adaptation.

Contour-based instance segmentation has been actively studied, thanks to its flexibility and elegance in processing visual objects within complex backgrounds. In this work, we propose a novel deep network architecture,i.e., PolySnake, for generic contour-based instance segmentation. Motivated by the classic Snake algorithm, the proposed PolySnake achieves superior and robust segmentation performance with an iterative and progressive contour refinement strategy. Technically, PolySnake introduces a recurrent update operator to estimate the object contour iteratively. It maintains a single estimate of the contour that is progressively deformed toward the object boundary. At each iteration, PolySnake builds a semantic-rich representation for the current contour and feeds it to the recurrent operator for further contour adjustment. Through the iterative refinements, the contour progressively converges to a stable status that tightly encloses the object instance. Beyond the scope of general instance segmentation, extensive experiments are conducted to validate the effectiveness and generalizability of our PolySnake in two additional specific task scenarios, including scene text detection and lane detection. The results demonstrate that the proposed PolySnake outperforms the existing advanced methods on several multiple prevalent benchmarks across the three tasks. The codes and pre-trained models are available at https://github.com/fh2019ustc/PolySnake.

In recent years, tremendous efforts have been made on document image rectification, but existing advanced algorithms are limited to processing restricted document images, i.e., the input images must incorporate a complete document. Once the captured image merely involves a local text region, its rectification quality is degraded and unsatisfactory. Our previously proposed DocTr, a transformer-assisted network for document image rectification, also suffers from this limitation. In this work, we present DocTr++, a novel unified framework for document image rectification, without any restrictions on the input distorted images. Our major technical improvements can be concluded in three aspects. Firstly, we upgrade the original architecture by adopting a hierarchical encoder-decoder structure for multi-scale representation extraction and parsing. Secondly, we reformulate the pixel-wise mapping relationship between the unrestricted distorted document images and the distortion-free counterparts. The obtained data is used to train our DocTr++ for unrestricted document image rectification. Thirdly, we contribute a real-world test set and metrics applicable for evaluating the rectification quality. To our best knowledge, this is the first learning-based method for the rectification of unrestricted document images. Extensive experiments are conducted, and the results demonstrate the effectiveness and superiority of our method. We hope our DocTr++ will serve as a strong baseline for generic document image rectification, prompting the further advancement and application of learning-based algorithms.

In the ever-evolving domain of multimedia, the significance of multi-modality understanding cannot be overstated. As multimedia content becomes increasingly sophisticated and ubiquitous, the ability to effectively combine and analyze the diverse information from different types of data, such as text, audio, image, video and point clouds, will be paramount in pushing the boundaries of what technology can achieve in understanding and interacting with the world around us. Accordingly, multi-modality understanding has attracted a tremendous amount of research, establishing itself as an emerging topic. Pre-trained models, in particular, have revolutionized this field, providing a way to leverage vast amounts of data without task-specific annotation to facilitate various downstream tasks.

LiDAR and Radar are two complementary sensing approaches in that LiDAR specializes in capturing an object's 3D shape while Radar provides longer detection ranges as well as velocity hints. Though seemingly natural, how to efficiently combine them for improved feature representation is still unclear. The main challenge arises from that Radar data are extremely sparse and lack height information. Therefore, directly integrating Radar features into LiDAR-centric detection networks is not optimal. In this work, we introduce a bi-directional LiDAR-Radar fusion framework, termed Bi-LRFusion, to tackle the challenges and improve 3D detection for dynamic objects. Technically, Bi-LRFusion involves two steps: first, it enriches Radar's local features by learning important details from the LiDAR branch to alleviate the problems caused by the absence of height information and extreme sparsity; second, it combines LiDAR features with the enhanced Radar features in a unified bird's-eye-view representation. We conduct extensive experiments on nuScenes and ORR datasets, and show that our Bi-LRFusion achieves state-of-the-art performance for detecting dynamic objects. Notably, Radar data in these two datasets have different formats, which demonstrates the generalizability of our method. Codes will be published.

Automatic recognition of disordered and elderly speech remains a highly challenging task to date due to the difficulty in collecting such data in large quantities. This paper explores a series of approaches to integrate domain adapted Self-Supervised Learning (SSL) pre-trained models into TDNN and Conformer ASR systems for dysarthric and elderly speech recognition: a) input feature fusion between standard acoustic frontends and domain adapted wav2vec2.0 speech representations; b) frame-level joint decoding of TDNN systems separately trained using standard acoustic features alone and with additional wav2vec2.0 features; and c) multi-pass decoding involving the TDNN/Conformer system outputs to be rescored using domain adapted wav2vec2.0 models. In addition, domain adapted wav2vec2.0 representations are utilized in acoustic-to-articulatory (A2A) inversion to construct multi-modal dysarthric and elderly speech recognition systems. Experiments conducted on the UASpeech dysarthric and DementiaBank Pitt elderly speech corpora suggest TDNN and Conformer ASR systems integrated domain adapted wav2vec2.0 models consistently outperform the standalone wav2vec2.0 models by statistically significant WER reductions of 8.22% and 3.43% absolute (26.71% and 15.88% relative) on the two tasks respectively. The lowest published WERs of 22.56% (52.53% on very low intelligibility, 39.09% on unseen words) and 18.17% are obtained on the UASpeech test set of 16 dysarthric speakers, and the DementiaBank Pitt test set respectively.

Automatic recognition of disordered speech remains a highly challenging task to date. The underlying neuro-motor conditions, often compounded with co-occurring physical disabilities, lead to the difficulty in collecting large quantities of impaired speech required for ASR system development. This paper presents novel variational auto-encoder generative adversarial network (VAE-GAN) based personalized disordered speech augmentation approaches that simultaneously learn to encode, generate and discriminate synthesized impaired speech. Separate latent features are derived to learn dysarthric speech characteristics and phoneme context representations. Self-supervised pre-trained Wav2vec 2.0 embedding features are also incorporated. Experiments conducted on the UASpeech corpus suggest the proposed adversarial data augmentation approach consistently outperformed the baseline speed perturbation and non-VAE GAN augmentation methods with trained hybrid TDNN and End-to-end Conformer systems. After LHUC speaker adaptation, the best system using VAE-GAN based augmentation produced an overall WER of 27.78% on the UASpeech test set of 16 dysarthric speakers, and the lowest published WER of 57.31% on the subset of speakers with "Very Low" intelligibility.

Early diagnosis of Alzheimer’s disease (AD) is crucial in facilitating preventive care and to delay further progression. Speech based automatic AD screening systems provide a non-intrusive and more scalable alternative to other clinical screening techniques. Textual embedding features produced by pre-trained language models (PLMs) such as BERT are widely used in such systems. However, PLM domain fine-tuning is commonly based on the masked word or sentence prediction costs that are inconsistent with the back-end AD detection task. To this end, this paper investigates the use of prompt-based fine-tuning of PLMs that consistently uses AD classification errors as the training objective function. Disfluency features based on hesitation or pause filler token frequencies are further incorporated into prompt phrases during PLM fine-tuning. The decision voting based combination among systems using different PLMs (BERT and RoBERTa) or systems with different fine-tuning paradigms (conventional masked-language modelling fine-tuning and prompt-based fine-tuning) is further applied. Mean, standard deviation (std) and the maximum among accuracy scores over 15 experiment runs are adopted as performance measurements for the AD detection system. Mean detection accuracy of 84.20% (with std 2.09%, best 87.5%) and 82.64% (with std 4.0%, best 89.58%) were obtained using manual and ASR speech transcripts respectively on the ADReSS20 test set consisting of 48 elderly speakers.

In fisheye images, rich distinct distortion patterns are regularly distributed in the image plane. These distortion patterns are independent of the visual content and provide informative cues for rectification. To make the best of such rectification cues, we introduce SimFIR, a simple framework for fisheye image rectification based on self-supervised representation learning. Technically, we first split a fisheye image into multiple patches and extract their representations with a Vision Transformer (ViT). To learn fine-grained distortion representations, we then associate different image patches with their specific distortion patterns based on the fisheye model, and further subtly design an innovative unified distortion-aware pretext task for their learning. The transfer performance on the downstream rectification task is remarkably boosted, which verifies the effectiveness of the learned representations. Extensive experiments are conducted, and the quantitative and qualitative results demonstrate the superiority of our method over the state-of-the-art algorithms as well as its strong generalization ability on real-world fisheye images.

In 3D human action recognition, limited supervised data makes it challenging to fully tap into the modeling potential of powerful networks such as transformers. As a result, researchers have been actively investigating effective self-supervised pre-training strategies. In this work, we show that instead of following the prevalent pretext task to perform masked self-component reconstruction in human joints, explicit contextual motion modeling is key to the success of learning effective feature representation for 3D action recognition. Formally, we propose the Masked Motion Prediction (MAMP) framework. To be specific, the proposed MAMP takes as input the masked spatio-temporal skeleton sequence and predicts the corresponding temporal motion of the masked human joints. Considering the high temporal redundancy of the skeleton sequence, in our MAMP, the motion information also acts as an empirical semantic richness prior that guide the masking process, promoting better attention to semantically rich temporal regions. Extensive experiments on NTU-60, NTU-120, and PKU-MMD datasets show that the proposed MAMP pre-training substantially improves the performance of the adopted vanilla transformer, achieving state-of-the-art results without bells and whistles. The source code of our MAMP is available at https://github.com/maoyunyao/MAMP.

Transformer-based methods have swept the benchmarks on 2D and 3D detection on images. Because tokenization before the attention mechanism drops the spatial information, positional encoding becomes critical for those methods. Recent works found that encodings based on samples of the 3D viewing rays can significantly improve the quality of multi-camera 3D object detection. We hypothesize that 3D point locations can provide more information than rays. Therefore, we introduce 3D point positional encoding, 3DPPE, to the 3D detection Transformer decoder. Although 3D measurements are not available at the inference time of monocular 3D object detection, 3DPPE uses predicted depth to approximate the real point positions. Our hybrid-depth module combines direct and categorical depth to estimate the refined depth of each pixel. Despite the approximation, 3DPPE achieves 46.0 mAP and 51.4 NDS on the competitive nuScenes dataset, significantly outperforming encodings based on ray samples. The code is available at https://github.com/drilistbox/3DPPE.

We tackle the data scarcity challenge in few-shot point cloud recognition of 3D objects by using a joint prediction from a conventional 3D model and a well-trained 2D model. Surprisingly, such an ensemble, though seems trivial, has hardly been shown effective in recent 2D-3D models. We find out the crux is the less effective training for the "joint hard samples", which have high confidence prediction on different wrong labels, implying that the 2D and 3D models do not collaborate well. To this end, our proposed invariant training strategy, called INVJOINT, does not only emphasize the training more on the hard samples, but also seeks the invariance between the conflicting 2D and 3D ambiguous predictions. INVJOINT can learn more collaborative 2D and 3D representations for better ensemble. Extensive experiments on 3D shape classification with widely adopted ModelNet10/40, ScanObjectNN and Toys4K, and shape retrieval with ShapeNet-Core validate the superiority of our INVJOINT. Codes will be publicly Available1.

Recent progress in weakly supervised object detection is featured by a combination of multiple instance detection networks (MIDN) and ordinal online refinement. However, with only image-level annotation, MIDN inevitably assigns high scores to some unexpected region proposals when generating pseudo labels. These inaccurate high-scoring region proposals will mislead the training of subsequent refinement modules and thus hamper the detection performance. In this work, we explore how to ameliorate the quality of pseudo-labeling in MIDN. Formally, we devise Cyclic-Bootstrap Labeling (CBL), a novel weakly supervised object detection pipeline, which optimizes MIDN with rank information from a reliable teacher network. Specifically, we obtain this teacher network by introducing a weighted exponential moving average strategy to take advantage of various refinement modules. A novel class-specific ranking distillation algorithm is proposed to leverage the output of weighted ensembled teacher network for distilling MIDN with rank information. As a result, MIDN is guided to assign higher scores to accurate proposals among their neighboring ones, thus benefiting the subsequent pseudo labeling. Extensive experiments on the prevalent PASCAL VOC 2007 & 2012 and COCO datasets demonstrate the superior performance of our CBL framework. Code will be available at https://github.com/Yinyf0804/WSOD-CBL/.

Zero-shot Human-Object Interaction (HOI) detection aims to identify both seen and unseen HOI categories. A strong zero-shot HOI detector is supposed to be not only capable of discriminating novel interactions but also robust to positional distribution discrepancy between seen and unseen categories when locating human-object pairs. However, top-performing zero-shot HOI detectors rely on seen and predefined unseen categories to distill knowledge from CLIP and jointly locate human-object pairs without considering the potential positional distribution discrepancy, leading to impaired transferability. In this paper, we introduce CLIP4HOI, a novel framework for zero-shot HOI detection. CLIP4HOI is developed on the vision-language model CLIP and ameliorates the above issues in the following two aspects. First, to avoid the model from overfitting to the joint positional distribution of seen human-object pairs, we seek to tackle the problem of zero-shot HOI detection in a disentangled two-stage paradigm. To be specific, humans and objects are independently identified and all feasible human-object pairs are processed by Human-Object interactor for pairwise proposal generation. Second, to facilitate better transferability, the CLIP model is elaborately adapted into a fine-grained HOI classifier for proposal discrimination, avoiding data-sensitive knowledge distillation. Finally, experiments on prevalent benchmarks show that our CLIP4HOI outperforms previous approaches on both rare and unseen categories, and sets a series of state-of-the-art records under a variety of zero-shot settings.

Currently, LiDAR-based 3D detectors are broadly categorized into two groups, namely, BEV-based detectors and cluster-based detectors. BEV-based detectors capture the contextual information from the Bird's Eye View (BEV) and fill their center voxels via feature diffusion with a stack of convolution layers, which, however, weakens the capability of presenting an object with the center point. On the other hand, cluster-based detectors exploit the voting mechanism and aggregate the foreground points into object-centric clusters for further prediction. In this paper, we explore how to effectively combine these two complementary representations into a unified framework. Specifically, we propose a new 3D object detection framework, referred to as CluB, which incorporates an auxiliary cluster-based branch into the BEV-based detector by enriching the object representation at both feature and query levels. Technically, CluB is comprised of two steps. First, we construct a cluster feature diffusion module to establish the association between cluster features and BEV features in a subtle and adaptive fashion. Based on that, an imitation loss is introduced to distill object-centric knowledge from the cluster features to the BEV features. Second, we design a cluster query generation module to leverage the voting centers directly from the cluster branch, thus enriching the diversity of object queries. Meanwhile, a direction loss is employed to encourage a more accurate voting center for each cluster. Extensive experiments are conducted on Waymo and nuScenes datasets, and our CluB achieves state-of-the-art performance on both benchmarks.

Abstract 3D object detection is receiving increasing attention from both industry and academia thanks to its wide applications in various fields. In this paper, we propose Point-Voxel Region-based Convolution Neural Networks (PV-RCNNs) for 3D object detection on point clouds. First, we propose a novel 3D detector, PV-RCNN, which boosts the 3D detection performance by deeply integrating the feature learning of both point-based set abstraction and voxel-based sparse convolution through two novel steps, i.e. , the voxel-to-keypoint scene encoding and the keypoint-to-grid RoI feature abstraction. Second, we propose an advanced framework, PV-RCNN++, for more efficient and accurate 3D object detection. It consists of two major improvements: sectorized proposal-centric sampling for efficiently producing more representative keypoints, and VectorPool aggregation for better aggregating local point features with much less resource consumption. With these two strategies, our PV-RCNN++ is about $$3\times $$ 3 × faster than PV-RCNN, while also achieving better performance. The experiments demonstrate that our proposed PV-RCNN++ framework achieves state-of-the-art 3D detection performance on the large-scale and highly-competitive Waymo Open Dataset with 10 FPS inference speed on the detection range of $$150m \times 150m$$ 150 m × 150 m .

In this work, we explore neat yet effective Transformer-based frameworks for visual grounding. The previous methods generally address the core problem of visual grounding, i.e., multi-modal fusion and reasoning, with manually-designed mechanisms. Such heuristic designs are not only complicated but also make models easily overfit specific data distributions. To avoid this, we first propose TransVG, which establishes multi-modal correspondences by Transformers and localizes referred regions by directly regressing box coordinates. We empirically show that complicated fusion modules can be replaced by a simple stack of Transformer encoder layers with higher performance. However, the core fusion Transformer in TransVG is stand-alone against uni-modal encoders, and thus should be trained from scratch on limited visual grounding data, which makes it hard to be optimized and leads to sub-optimal performance. To this end, we further introduce TransVG++ to make two-fold improvements. For one thing, we upgrade our framework to a purely Transformer-based one by leveraging Vision Transformer (ViT) for vision feature encoding. For another, we devise Language Conditioned Vision Transformer that removes external fusion modules and reuses the uni-modal ViT for vision-language fusion at the intermediate layers. We conduct extensive experiments on five prevalent datasets, and report a series of state-of-the-art records.

In pixel-based reinforcement learning (RL), the states are raw video frames, which are mapped into hidden representation before feeding to a policy network. To improve sample efficiency of state representation learning, recently, the most prominent work is based on contrastive unsupervised representation. Witnessing that consecutive video frames in a game are highly correlated, to further improve data efficiency, we propose a new algorithm, i.e., masked contrastive representation learning for RL (M-CURL), which takes the correlation among consecutive inputs into consideration. In our architecture, besides a CNN encoder for hidden presentation of input state and a policy network for action selection, we introduce an auxiliary Transformer encoder module to leverage the correlations among video frames. During training, we randomly mask the features of several frames, and use the CNN encoder and Transformer to reconstruct them based on context frames. The CNN encoder and Transformer are jointly trained via contrastive learning where the reconstructed features should be similar to the ground-truth ones while dissimilar to others. During policy evaluation, the CNN encoder and the policy network are used to take actions, and the Transformer module is discarded. Our method achieves consistent improvements over CURL on 14 out of 16 environments from DMControl suite and 23 out of 26 environments from Atari 2600 Games. The code is available at https://github.com/teslacool/m-curl.

Speaker adaptation techniques provide a powerful solution to customise automatic speech recognition (ASR) systems for individual users. Practical application of unsupervised model-based speaker adaptation techniques to data intensive end-to-end ASR systems is hindered by the scarcity of speaker-level data and performance sensitivity to transcription errors. To address these issues, a set of compact and data efficient speaker-dependent (SD) parameter representations are used to facilitate both speaker adaptive training and test-time unsupervised speaker adaptation of state-of-the-art Conformer ASR systems. The sensitivity to supervision quality is reduced using a confidence score-based selection of the less erroneous subset of speaker-level adaptation data. Two lightweight confidence score estimation modules are proposed to produce more reliable confidence scores. The data sparsity issue, which is exacerbated by data selection, is addressed by modelling the SD parameter uncertainty using Bayesian learning. Experiments on the benchmark 300-hour Switchboard and the 233-hour AMI datasets suggest that the proposed confidence score-based adaptation schemes consistently outperformed the baseline speaker-independent (SI) Conformer model and conventional non-Bayesian, point estimate-based adaptation using no speaker data selection. Similar consistent performance improvements were retained after external Transformer and LSTM language model rescoring. In particular, on the 300-hour Switchboard corpus, statistically significant WER reductions of 1.0%, 1.3%, and 1.4% absolute (9.5%, 10.9%, and 11.3% relative) were obtained over the baseline SI Conformer on the NIST Hub5’00, RT02, and RT03 evaluation sets respectively. Similar WER reductions of 2.7% and 3.3% absolute (8.9% and 10.2% relative) were also obtained on the AMI development and evaluation sets.

Accurate recognition of cocktail party speech containing overlapping speakers, noise and reverberation remains a highly challenging task to date. Motivated by the invariance of visual modality to acoustic signal corruption, an audio-visual multi-channel speech separation, dereverberation and recognition approach featuring a full incorporation of visual information into all system components is proposed in this paper. The efficacy of the video input is consistently demonstrated in mask-based MVDR speech separation, DNN-WPE or spectral mapping (SpecM) based speech dereverberation front-end and Conformer ASR back-end. Audio-visual integrated front-end architectures performing speech separation and dereverberation in a pipelined or joint fashion via mask-based WPD are investigated. The error cost mismatch between the speech enhancement front-end and ASR back-end components is minimized by end-to-end jointly fine-tuning using either the ASR cost function alone, or its interpolation with the speech enhancement loss. Experiments were conducted on the mixture overlapped and reverberant speech data constructed using simulation or replay of the Oxford LRS2 dataset. The proposed audio-visual multi-channel speech separation, dereverberation and recognition systems consistently outperformed the comparable audio-only baseline by 9.1% and 6.2% absolute (41.7% and 36.0% relative) word error rate (WER) reductions. Consistent speech enhancement improvements were also obtained on PESQ, STOI and SRMR scores.

Motion prediction from raw LiDAR sensor data has drawn increasing attention and led to a surge of studies following two main paradigms. One paradigm is global motion paradigm, which simultaneously detects objects from point clouds and predicts the trajectories of each object in the future. The other paradigm is local motion paradigm, which directly performs dense motion prediction pointwisely. We observe that global motion prediction can benefit from local motion representation, since it contains rich local displacement contexts that are not explicitly exploited in global motion prediction. Correspondingly, local motion prediction can benefit from global motion representation, since it provides object contexts to improve prediction consistency inside an object. However, the complement of these two motion representations has not fully explored in the literature. To this end, we propose Hybrid Motion Representation Learning (HyMo), a unified framework to address the problem of motion prediction by making the best of both global and local motion cues. We have conducted extensive experiments on nuScenes dataset. The experimental results demonstrate that the learned hybrid motion representation achieves state-of-the-art performance on both global and local motion prediction tasks.

Existing solutions for weakly supervised object detection (WSOD) generally follow the multiple instance learning (MIL) paradigm to formulate WSOD as a multi-class classification problem over a set of region proposals. However, without the supervision signal of ground-truth boxes, the training objective of multi-class classification makes the detectors devote main efforts to finding the most common pattern of each class, as the common pattern is always the most discriminative evidence for classification. In addition, although learning from distinguishing multiple foreground classes, the detectors can still ignore to differentiate foreground regions from the background ones, which causes false alarm in prediction. These two points account for the limited localization capability of MIL-based WSOD methods. To this end, we propose foreground information guided WSOD (FI-WSOD), a novel framework that introduces an extra foreground-background binary classification (F-BBC) sub-task to the original MIL-based WSOD paradigm. At the training stage, the involvement of F-BBC task not only improves the feature representation of the network, but also provides extra information from the foreground-background perspective. By leveraging the learnt foreground information, a Foreground Guided Self-Training (FGST) module is further proposed to filter out noisy samples, and to mine representative seeds from the remaining proposals. Moreover, a Multi-Seed Training strategy is performed to reduce the impact of noisy labels when training the self-training networks in FGST. We have conducted extensive experiments on the prevalent Pascal VOC 2007, Pascal VOC 2012 and MSCOCO datasets, and report a series of state-of-the-art records achieved by our proposed framework.

It has been well recognized that fusing the complementary information from depth-aware LiDAR point clouds and semantic-rich stereo images would benefit 3D object detection. Nevertheless, it is non-trivial to explore the inherently unnatural interaction between sparse 3D points and dense 2D pixels. To ease this difficulty, the recent approaches generally project the 3D points onto the 2D image plane to sample the image data and then aggregate the data at the points. However, these approaches often suffer from the mismatch between the resolution of point clouds and RGB images, leading to sub-optimal performance. Specifically, taking the sparse points as the multi-modal data aggregation locations causes severe information loss for high-resolution images, which in turn undermines the effectiveness of multi-sensor fusion. In this paper, we presentVPFNet—a new architecture that cleverly aligns and aggregates the point cloud and image data at the “virtual” points. Particularly, with their density lying between that of the 3D points and 2D pixels, the virtual points can nicely bridge the resolution gap between the two sensors, and thus preserve more information for processing. Moreover, we also investigate the data augmentation techniques that can be applied to both point clouds and RGB images, as the data augmentation has made non-negligible contribution towards 3D object detectors to date. We have conducted extensive experiments on KITTI dataset, and have observed good performance compared to the state-of-the-art methods. Remarkably, ourVPFNetachieves 83.21% moderate$AP_{3D}$and 91.86% moderate$AP_{BEV}$on the KITTI test set. The network design also takes computation efficiency into consideration – we can achieve a FPS of 15 on a single NVIDIA RTX 2080Ti GPU.

Despite the rapid advance of automatic speech recognition (ASR) technologies, accurate recognition of cocktail party speech characterised by the interference from overlapping speakers, background noise and room reverberation remains a highly challenging task to date. Motivated by the invariance of visual modality to acoustic signal corruption, audio-visual speech enhancement techniques have been developed, although predominantly targeting overlapping speech separation and recognition tasks. In this paper, an audio-visual multi-channel speech separation, dereverberation and recognition approach featuring a full incorporation of visual information into all three stages of the system is proposed. The advantage of the additional visual modality over using audio only is demonstrated on two neural dereverberation approaches based on DNN-WPE and spectral mapping respectively. The learning cost function mismatch between the separation and dereverberation models and their integration with the back-end recognition system is minimised using fine-tuning on the MSE and LF-MMI criteria. Experiments conducted on the LRS2 dataset suggest that the proposed audio-visual multi-channel speech separation, dereverberation and recognition system outperforms the baseline audio-visual multi-channel speech separation and recognition system containing no dereverberation module by a statistically significant word error rate (WER) reduction of 2.06 % absolute (8.77 % relative).

Pulmonary nodule classification from computerized tomography(CT) Scans is a vital task for the early screening of Lung cancers. The algorithm is aiming at distinguishing malignant pulmonary nodules, benign nodules and the ones with their subtypes. In this paper, we defined a detailed pulmonary nodule classification task considering 5 semantic labels. We are facing with a series of non-trival problems dealing with such a task. First, the available medical image data for training is quite limited. We enlarged the training dataset by cropping out three-dimension(3D) volume of each pulmonary nodule and generating 15 planes with different orientations from these volumes. Secondly, the global modeling ability of the existing convolutional neural network(CNN) based architectures can not meet the need of medical image analysis well. To learn discriminative abstract information, we down-sample feature maps between successive stages and adopt the BotNet-50 backbone which is a combination of ResNet backbone and self-attention modules. Such an architecture can extract local and non-local information in low-level and high-level layers, respectively. Last but not the least, the data distribution of training data and testing data don't share similar distribution in real-world multi-center medical image classification scenes. We assigned the samples with modified wights while calculating the loss value for optimization. The proposed method can eliminate the spurious correlation between features and labels. Experiments demonstrate the effectiveness of each component.

Fundamental modelling differences between hybrid and end-to-end (E2E) automatic speech recognition (ASR) systems create large diversity and complementarity among them. This paper investigates multi-pass rescoring and cross adaptation based system combination approaches for hybrid TDNN and Conformer E2E ASR systems. In multi-pass rescoring, state-of-the-art hybrid LF-MMI trained CNN-TDNN system featuring speed perturbation, SpecAugment and Bayesian learning hidden unit contributions (LHUC) speaker adaptation was used to produce initial N-best outputs before being rescored by the speaker adapted Conformer system using a 2-way cross system score interpolation. In cross adaptation, the hybrid CNN-TDNN system was adapted to the 1-best output of the Conformer system or vice versa. Experiments on the 300-hour Switchboard corpus suggest that the combined systems derived using either of the two system combination approaches outperformed the individual systems. The best combined system obtained using multi-pass rescoring produced statistically significant word error rate (WER) reductions of 2.5% to 3.9% absolute (22.5% to 28.9% relative) over the stand alone Conformer system on the NIST Hub5'00, Rt03 and Rt02 evaluation data.

A key challenge for automatic speech recognition (ASR) systems is to model the speaker level variability.In this paper, compact speaker dependent learning hidden unit contributions (LHUC) are used to facilitate both speaker adaptive training (SAT) and test time unsupervised speaker adaptation for stateof-the-art Conformer based end-to-end ASR systems.The sensitivity during adaptation to supervision error rate is reduced using confidence score based selection of the more "trustworthy" subset of speaker specific data.A confidence estimation module is used to smooth the over-confident Conformer decoder output probabilities before serving as confidence scores.The increased data sparsity due to speaker level data selection is addressed using Bayesian estimation of LHUC parameters.Experiments on the 300-hour Switchboard corpus suggest that the proposed LHUC-SAT Conformer with confidence score based test time unsupervised adaptation outperformed the baseline speaker independent and i-vector adapted Conformer systems by up to 1.0%, 1.0%, and 1.2% absolute (9.0%, 7.9%, and 8.9% relative) word error rate (WER) reductions on the NIST Hub5'00, RT02, and RT03 evaluation sets respectively.Consistent performance improvements were retained after external Transformer and LSTM language models were used for rescoring.

Early diagnosis of Alzheimer's disease (AD) is crucial in facilitating preventive care to delay further progression. This paper presents the development of a state-of-the-art Conformer based speech recognition system built on the DementiaBank Pitt corpus for automatic AD detection. The baseline Conformer system trained with speed perturbation and SpecAugment based data augmentation is significantly improved by incorporating a set of purposefully designed modeling features, including neural architecture search based auto-configuration of domain-specific Conformer hyper-parameters in addition to parameter fine-tuning; fine-grained elderly speaker adaptation using learning hidden unit contributions (LHUC); and two-pass cross-system rescoring based combination with hybrid TDNN systems. An overall word error rate (WER) reduction of 13.6% absolute (34.8% relative) was obtained on the evaluation data of 48 elderly speakers. Using the final systems' recognition outputs to extract textual features, the best-published speech recognition based AD detection accuracy of 91.7% was obtained.

BACKGROUND: Patients in the intensive care unit (ICU) are often in critical condition and have a high mortality rate. Accurately predicting the survival probability of ICU patients is beneficial to timely care and prioritizing medical resources to improve the overall patient population survival. Models developed by deep learning (DL) algorithms show good performance on many models. However, few DL algorithms have been validated in the dimension of survival time or compared with traditional algorithms. METHODS: Variables from the Early Warning Score, Sequential Organ Failure Assessment Score, Simplified Acute Physiology Score II, Acute Physiology and Chronic Health Evaluation (APACHE) II, and APACHE IV models were selected for model development. The Cox regression, random survival forest (RSF), and DL methods were used to develop prediction models for the survival probability of ICU patients. The prediction performance was independently evaluated in the MIMIC-III Clinical Database (MIMIC-III), the eICU Collaborative Research Database (eICU), and Shanghai Pulmonary Hospital Database (SPH). RESULTS: Forty variables were collected in total for model development. 83 943 participants from 3 databases were included in the study. The New-DL model accurately stratified patients into different survival probability groups with a C-index of >0.7 in the MIMIC-III, eICU, and SPH, performing better than the other models. The calibration curves of the models at 3 and 10 days indicated that the prediction performance was good. A user-friendly interface was developed to enable the model's convenience. CONCLUSIONS: Compared with traditional algorithms, DL algorithms are more accurate in predicting the survival probability during ICU hospitalization. This novel model can provide reliable, individualized survival probability prediction.

State-of-the-art automatic speech recognition (ASR) system development is data and computation intensive. The optimal design of deep neural networks (DNNs) for these systems often require expert knowledge and empirical evaluation. In this paper, a range of neural architecture search (NAS) techniques are used to automatically learn two types of hyper-parameters of factored time delay neural networks (TDNN-Fs): i) the left and right splicing context offsets; and ii) the dimensionality of the bottleneck linear projection at each hidden layer. These techniques include the differentiable neural architecture search (DARTS) method integrating architecture learning with lattice-free MMI training; Gumbel-Softmax and pipelined DARTS methods reducing the confusion over candidate architectures and improving the generalization of architecture selection; and Penalized DARTS incorporating resource constraints to balance the trade-off between performance and system complexity. Parameter sharing among TDNN-F architectures allows an efficient search over up to$7^{28}$different systems. Statistically significant word error rate (WER) reductions of up to 1.2% absolute and relative model size reduction of 31% were obtained over a state-of-the-art 300-hour Switchboard corpus trained baseline LF-MMI TDNN-F system featuring speed perturbation, i-Vector and learning hidden unit contribution (LHUC) based speaker adaptation as well as RNNLM rescoring. Performance contrasts on the same task against recent end-to-end systems reported in the literature suggest the best NAS auto-configured system achieves state-of-the-art WERs of 9.9% and 11.1% on the NIST Hub5’ 00 and Rt03 s test sets respectively with up to 96% model size reduction. Further analysis using Bayesian learning shows that the proposed NAS approaches can effectively minimize the structural redundancy in the TDNN-F systems and reduce their model parameter uncertainty. Consistent performance improvements were also obtained on a UASpeech dysarthric speech recognition task.

Temporal language grounding (TLG) is a fundamental and challenging problem for vision and language understanding. Existing methods mainly focus on fully supervised setting with temporal boundary labels for training, which, however, suffers expensive cost of annotation. In this work, we are dedicated to weakly supervised TLG, where multiple description sentences are given to an untrimmed video without temporal boundary labels. In this task, it is critical to learn a strong cross-modal semantic alignment between sentence semantics and visual content. To this end, we introduce a novel weakly supervised temporal adjacent network (WSTAN) for temporal language grounding. Specifically, WSTAN learns cross-modal semantic alignment by exploiting temporal adjacent network in a multiple instance learning (MIL) paradigm, with a whole description paragraph as input. Moreover, we integrate a complementary branch into the framework, which explicitly refines the predictions with pseudo supervision from the MIL stage. An additional self-discriminating loss is devised on both the MIL branch and the complementary branch, aiming to enhance semantic discrimination by self-supervising. Extensive experiments are conducted on three widely used benchmark datasets,i.e., ActivityNet-Captions, Charades-STA, and DiDeMo, and the results demonstrate the effectiveness of our approach.

Recent advances on 3D object detection heavily rely on how the 3D data are represented, i.e., voxel-based or point-based representation. Many existing high performance 3D detectors are point-based because this structure can better retain precise point positions. Nevertheless, point-level features lead to high computation overheads due to unordered storage. In contrast, the voxel-based structure is better suited for feature extraction but often yields lower accuracy because the input data are divided into grids. In this paper, we take a slightly different viewpoint --- we find that precise positioning of raw points is not essential for high performance 3D object detection and that the coarse voxel granularity can also offer sufficient detection accuracy. Bearing this view in mind, we devise a simple but effective voxel-based framework, named Voxel R-CNN. By taking full advantage of voxel features in a two-stage approach, our method achieves comparable detection accuracy with state-of-the-art point-based models, but at a fraction of the computation cost. Voxel R-CNN consists of a 3D backbone network, a 2D bird-eye-view (BEV) Region Proposal Network, and a detect head. A voxel RoI pooling is devised to extract RoI features directly from voxel features for further refinement. Extensive experiments are conducted on the widely used KITTI Dataset and the more recent Waymo Open Dataset. Our results show that compared to existing voxel-based methods, Voxel R-CNN delivers a higher detection accuracy while maintaining a real-time frame processing rate, i.e., at a speed of 25 FPS on an NVIDIA RTX 2080 Ti GPU. The code is available at https://github.com/djiajunustc/Voxel-R-CNN.

Recent progress on weakly supervised object detection (WSOD) is characterized by formulating WSOD as a Multiple Instance Learning (MIL) problem and taking online refinement with the selected region proposals from MIL. However, MIL inclines to select the most discriminative part rather than the entire instance as the top-scoring region proposals, which leads to weak localization capability for weakly supervised object detectors. We attribute this problem to the limited intra-class diversity within a single image. Specifically, due to the lack of annotated bounding boxes, the network tends to focus on the most common parts of each class and neglect the diverse parts of objects. To solve the problem, we introduce a novel Instance Mining with Class Feature Banks (IM-CFB) framework, which includes a Class Feature Banks (CFB) module and a Feature Guided Instance Mining (FGIM) algorithm. Concretely, Class Feature Banks (CFB) consist of sub-banks for each class, which are utilized to collect diversity information from a broader view. At the training stage, the RoI features of reliable region proposals are recorded and updated in the CFB. Then, FGIM leverages the features recorded in the CFB to ameliorate the region proposal selection of the MIL branch. Extensive experiments conducted on two publicly available datasets, Pascal VOC 2007 and 2012, demonstrate the effectiveness of our method. More remarkably, our method achieves 54.3% on mAP and 70.7% on CorLoc on Pascal VOC 2007. When further re-trained by a Fast-RCNN detector, we obtain to-date the best reported mAP and CorLoc of 55.8% and 72.2%, respectively.

In this paper, we present a neat yet effective transformer-based framework for visual grounding, namely TransVG, to address the task of grounding a language query to the corresponding region onto an image. The state-of-the-art methods, including two-stage or one-stage ones, rely on a complex module with manually-designed mechanisms to perform the query reasoning and multi-modal fusion. However, the involvement of certain mechanisms in fusion module design, such as query decomposition and image scene graph, makes the models easily overfit to datasets with specific scenarios, and limits the plenitudinous interaction between the visual-linguistic context. To avoid this caveat, we propose to establish the multi-modal correspondence by leveraging transformers, and empirically show that the complex fusion modules (e.g., modular attention network, dynamic graph, and multi-modal tree) can be replaced by a simple stack of transformer encoder layers with higher performance. Moreover, we re-formulate the visual grounding as a direct coordinates regression problem and avoid making predictions out of a set of candidates (i.e., region proposals or anchor boxes). Extensive experiments are conducted on five widely used datasets, and a series of state-of-the-art records are set by our TransVG. We build the benchmark of transformer-based visual grounding framework and make the code available at https://github.com/djiajunustc/TransVG.

As cameras are increasingly deployed in new application domains such as autonomous driving, performing 3D object detection on monocular images becomes an important task for visual scene understanding. Recent advances on monocular 3D object detection mainly rely on the "pseudo-LiDAR'' generation, which performs monocular depth estimation and lifts the 2D pixels to pseudo 3D points. However, depth estimation from monocular images, due to its poor accuracy, leads to inevitable position shift of pseudo-LiDAR points within the object. Therefore, the predicted bounding boxes may suffer from inaccurate location and deformed shape. In this paper, we present a novel neighbor-voting method that incorporates neighbor predictions to ameliorate object detection from severely deformed pseudo-LiDAR point clouds. Specifically, each feature point around the object forms their own predictions, and then the "consensus'' is achieved through voting. In this way, we can effectively combine the neighbors' predictions with local prediction and achieve more accurate 3D detection. To further enlarge the difference between the foreground region of interest (ROI) pseudo-LiDAR points and the background points, we also encode the ROI prediction scores of 2D foreground pixels into the corresponding pseudo-LiDAR points. We conduct extensive experiments on the KITTI benchmark to validate the merits of our proposed method. Our results on the bird's eye view detection outperform the state-of-the-art performance, especially for the "hard" level detection. The code is available at https://github.com/cxmomo/Neighbor-Vote.

In this work, we propose a new framework, called Document Image Transformer (DocTr), to address the issue of geometry and illumination distortion of the document images. Specifically, DocTr consists of a geometric unwarping transformer and an illumination correction transformer. By setting a set of learned query embedding, the geometric unwarping transformer captures the global context of the document image by self-attention mechanism and decodes the pixel-wise displacement solution to correct the geometric distortion. After geometric unwarping, our illumination correction transformer further removes the shading artifacts to improve the visual quality and OCR accuracy. Extensive evaluations are conducted on several datasets, and superior results are reported against the state-of-the-art methods. Remarkably, our DocTr achieves $20.02%$ Character Error Rate (CER), a $15%$ absolute improvement over the state-of-the-art methods. Moreover, it also shows high efficiency on running time and parameter count.

As an emerging data modal with precise distance sensing, LiDAR point clouds have been placed great expectations on 3D scene understanding. However, point clouds are always sparsely distributed in the 3D space, and with unstructured storage, which makes it difficult to represent them for effective 3D object detection. To this end, in this work, we regard point clouds as hollow-3D data and propose a new architecture, namely Hallucinated Hollow-3D R-CNN (H23D R-CNN), to address the problem of 3D object detection. In our approach, we first extract the multi-view features by sequentially projecting the point clouds into the perspective view and the bird-eye view. Then, we hallucinate the 3D representation by a novel bilaterally guided multi-view fusion block. Finally, the 3D objects are detected via a box refinement module with a novel Hierarchical Voxel RoI Pooling operation. The proposed H23D R-CNN provides a new angle to take full advantage of complementary information in the perspective view and the bird-eye view with an efficient framework. We evaluate our approach on the public KITTI Dataset and Waymo Open Dataset. Extensive experiments demonstrate the superiority of our method over the state-of-the-art algorithms with respect to both effectiveness and efficiency. The code is available athttps://github.com/djiajunustc/H-23D_R-CNN.

Recent advances in video object detection have characterized the exploration of temporal coherence across frames to enhance object detector. Nevertheless, previous solutions either rely on additional inputs (e.g., optical flow) to guide feature aggregation, or complex post-processing to associate bounding boxes. In this paper, we introduce a simple but effective design that learns instance identifiers for instance association in a meta-learning paradigm, which requires no auxiliary inputs or post-processing. Specifically, we present Meta-Learnt Instance Identifier Networks (namely MINet) that novelly meta-learns instance identifiers to recognize identical instances across frames in a single forward-pass, leading to the robust online linking of instances. Technically, depending on the detection results of previous frames, we teach MINet to learn the weights of an instance identifier on the fly, which can be well applied to up-coming frames. Such meta-learning paradigm enables instance identifiers to be flexibly adapted to novel frames at inference. Furthermore, MINet writes/updates the detection results of previous instances into memory and reads from memory when performing inference to encourage temporal consistency for video object detection. Our MINet is appealing in the sense that it is pluggable to any object detection model. Extensive experiments on ImageNet VID dataset demonstrate the superiority of MINet. More remarkably, by integrating MINet into Faster R-CNN, we achieve 80.2% mAP on ImageNet VID dataset.

Single shot detectors that are potentially faster and simpler than two-stage detectors tend to be more applicable to object detection in videos. Nevertheless, the extension of such object detectors from image to video is not trivial especially when appearance deterioration exists in videos, e.g., motion blur or occlusion. A valid question is how to explore temporal coherence across frames for boosting detection. In this paper, we propose to address the problem by enhancing per-frame features through aggregation of neighboring frames. Specifically, we present Single Shot Video Object Detector (SSVD) - a new architecture that novelly integrates feature aggregation into a one-stage detector for object detection in videos. Technically, SSVD takes Feature Pyramid Network (FPN) as backbone network to produce multi-scale features. Unlike the existing feature aggregation methods, SSVD, on one hand, estimates the motion and aggregates the nearby features along the motion path, and on the other, hallucinates features by directly sampling features from the adjacent frames in a two-stream structure. Extensive experiments are conducted on ImageNet VID dataset, and competitive results are reported when comparing to state-of-the-art approaches. More remarkably, for $448 \times 448$ input, SSVD achieves 79.2% mAP on ImageNet VID, by processing one frame in 85 ms on an Nvidia Titan X Pascal GPU. The code is available at https://github.com/ddjiajun/SSVD.

As an important topic in the multimedia and computer vision fields, salient object detection has been researched for years. Recently, state-of-the-art performance has been witnessed with the aid of the fully convolutional networks (FCNs) and the various pyramid-like encoder-decoder frameworks. Starting from a common encoder-decoder architecture, we enhance a residual refinement network with feature purification for better saliency estimation. To this end, we improve the global knowledge streams with intermediate supervisions for global saliency estimation and design a specific feature subtraction module for residual learning, respectively. On the basis of the strengthened network, we also introduce an attribute encoding sub-network (AENet) with a grid aggregation block (GAB) to guide the final saliency predictor to obtain more accurate saliency maps. Furthermore, the network is trained with a novel constraint loss besides the traditional cross-entropy loss to yield the finer results. Extensive experiments on five public benchmarks show our method achieves better or comparable performance compared with previous state-of-the-art methods.

It has been well recognized that modeling object-to-object relations would be helpful for object detection. Nevertheless, the problem is not trivial especially when exploring the interactions between objects to boost video object detectors. The difficulty originates from the aspect that reliable object relations in a video should depend on not only the objects in the present frame but also all the supportive objects extracted over a long range span of the video. In this paper, we introduce a new design to capture the interactions across the objects in spatio-temporal context. Specifically, we present Relation Distillation Networks (RDN) - a new architecture that novelly aggregates and propagates object relation to augment object features for detection. Technically, object proposals are first generated via Region Proposal Networks (RPN). RDN then, on one hand, models object relation via multi-stage reasoning, and on the other, progressively distills relation through refining supportive object proposals with high objectness scores in a cascaded manner. The learnt relation verifies the efficacy on both improving object detection in each frame and box linking across frames. Extensive experiments are conducted on ImageNet VID dataset, and superior results are reported when comparing to state-of-the-art methods. More remarkably, our RDN achieves 81.8% and 83.2% mAP with ResNet-101 and ResNeXt-101, respectively. When further equipped with linking and rescoring, we obtain to-date the best reported mAP of 83.8% and 84.7%.

Nowadays, inspired by the behavior and neuronal architecture of human visual system (HVS), region of interest (ROI) detection methods are investigated by integrating several cognitive features. However, most of them are complex and time-consuming. To solve this problem, a video ROI extraction algorithm based on information theory and cognitive features is proposed in this paper. Based on the information theory, the spatial and temporal information are computed to measure the spatial and temporal content of video sequences respectively. Utilizing the visual features fusion strategy (VFFS), visual information is obtained from the combination of spatial and temporal information. Then, perceptual stimulus-response model (PSRM) is established to map the visual information to match the visual saliency. Regions with saliency scores, which are higher than a self-adaptive threshold, are regarded as ROIs. Experimental results show that the proposed scheme reduces 50% computation complexity, and can extract ROI more effective compared to Itti's method. Furthermore, the proposed ROI extraction scheme can be easily applied in the practical multimedia processing and pattern recognition system, such as video summarization.