Recently, prompt learning has achieved remarkable success in adapting pre-trained Vision-Language Models (VLMs) to downstream tasks such as image classification. However, its application to the downstream Image-Text Retrieval (ITR) task is more challenging. We find that the challenge lies in discriminating both fine-grained attributes and similar subcategories of the downstream data. To address this challenge, we propose Dual prompt Learning with Joint Category-Attribute Reweighting (DCAR), a novel dual-prompt learning framework to achieve precise image-text matching. The framework dynamically adjusts prompt vectors from both semantic and visual dimensions to improve the performance of CLIP on the downstream ITR task. Based on the prompt paradigm, DCAR jointly optimizes attribute and category features to enhance fine-grained representation learning. Specifically, (1) at the attribute level, it dynamically updates the weights of attribute descriptions based on text-image mutual information correlation; and (2) at the category level, it introduces negative samples from multiple perspectives with category-matching weighting to learn subcategory distinctions. To validate our method, we construct the Fine-class Described Retrieval Dataset (FDRD), which serves as a challenging benchmark for ITR in downstream data domains. It covers over 1,500 downstream fine categories and 230,000 image-caption pairs with detailed attribute annotations. Extensive experiments on FDRD demonstrate that DCAR achieves state-of-the-art performance over existing baselines. The code and data are available at https://github.com/wyf202322/DCAR.

Existing object detection models are mainly trained on large-scale labeled datasets. However, annotating data for novel aerial object classes is expensive since it is time-consuming and may require expert knowledge. Thus, it is desirable to study label-efficient object detection methods on aerial images. In this work, we propose a zero-shot method for aerial object detection named visual Description Regularization, or DescReg. Concretely, we identify the weak semantic-visual correlation of the aerial objects and aim to address the challenge with prior descriptions of their visual appearance. Instead of directly encoding the descriptions into class embedding space which suffers from the representation gap problem, we propose to infuse the prior inter-class visual similarity conveyed in the descriptions into the embedding learning. The infusion process is accomplished with a newly designed similarity-aware triplet loss which incorporates structured regularization on the representation space. We conduct extensive experiments with three challenging aerial object detection datasets, including DIOR, xView, and DOTA. The results demonstrate that DescReg significantly outperforms the state-of-the-art ZSD methods with complex projection designs and generative frameworks, e.g., DescReg outperforms best reported ZSD method on DIOR by 4.5 mAP on unseen classes and 8.1 in HM. We further show the generalizability of DescReg by integrating it into generative ZSD methods as well as varying the detection architecture. Codes will be released at https://github.com/zq-zang/DescReg.

Key-value (KV) caching is an important technique to accelerate the inference of large language models (LLMs), but incurs significant memory overhead.To compress the size of KV cache, existing methods often compromise precision or require extra data for calibration, limiting their practicality in LLM deployment.In this paper, we introduce DecoQuant, a novel data-free low-bit quantization technique based on tensor decomposition methods, to effectively compress KV cache.Our core idea is to adjust the outlier distribution of the original matrix by performing tensor decomposition, so that the quantization difficulties are migrated from the matrix to decomposed local tensors.Specially, we find that outliers mainly concentrate on small local tensors, while large tensors tend to have a narrower value range.Based on this finding, we propose to apply low-bit quantization to the large tensor, while maintaining high-precision representation for the small tensor.Furthermore, we utilize the proposed quantization method to compress the KV cache of LLMs to accelerate the inference and develop an efficient dequantization kernel tailored specifically for DecoQuant.Through extensive experiments, DecoQuant demonstrates remarkable efficiency gains, showcasing up to a ∼75% reduction in memory footprint while maintaining comparable generation quality.

When it comes to object detection tasks, YOLO stands out for its impressive speed and efficiency. Nonetheless, deploying YOLO on resource-constrained devices remains a challenge due to its substantial model size and memory requirements. The direct application of conventional binary quantization strategies to YOLO can result in significant accuracy degradation. A prevalent solution is to introduce floating-point shortcuts. However, the increased computational demand and parameter complexity associated with these shortcuts limit their practical deployment on hardware platforms for optimal acceleration. To solve this problem, we propose a binary neural network (BNN) for object detection called BOB-YOLO to achieve a balanced performance in terms of computational speed, model size, and detection accuracy. Our BOB-YOLO fully leverages module-wise latency (MWL) to supervise the latency of floating-point shortcut branches by that of 1-bit trunk branches. This supervision maximizes the information carried by the floating-point data flow in shortcuts while maintaining latency within the limits set by the 1-bit convolution branch, thereby improving parallel computational efficiency. We also introduce the Roofline Model to address these limitations by considering both computational complexity and parameter compression, ensuring high computational intensity. Additionally, we propose a performance evaluation metric Pd, which provides an intuitive description of the trade-off between speed and accuracy, aligning closely with the practical requirements of binary quantization strategies. Extensive experiments on the VOC and COCO datasets demonstrate the significant advantages of our method over state-of-the-art BNN methods.

Data-dependent hashing methods aim to learn hash functions from the pairwise or triplet relationships among the data, which often lead to low efficiency and low collision rate by only capturing the local distribution of the data. To solve the limitation, we propose central similarity, in which the hash codes of similar data pairs are encouraged to approach a common center and those of dissimilar pairs to converge to different centers. As a new global similarity metric, central similarity can improve the efficiency and retrieval accuracy of hash learning. By introducing a new concept, hash centers, we principally formulate the computation of the proposed central similarity metric, in which the hash centers refer to a set of points scattered in the Hamming space with a sufficient mutual distance between each other. To construct well-separated hash centers, we provide two efficient methods: 1) leveraging the Hadamard matrix and Bernoulli distributions to generate data-independent hash centers and 2) learning data-dependent hash centers from data representations. Based on the proposed similarity metric and hash centers, we propose central similarity quantization (CSQ) that optimizes the central similarity between data points with respect to their hash centers instead of optimizing the local similarity to generate a high-quality deep hash function. We also further improve the CSQ with data-dependent hash centers, dubbed as CSQ with learnable center (CSQLC). The proposed CSQ and CSQLC are generic and applicable to image and video hashing scenarios. We conduct extensive experiments on large-scale image and video retrieval tasks, and the proposed CSQ yields noticeably boosted retrieval performance, i.e., 3%-20% in mean average precision (mAP) over the previous state-of-the-art methods, which also demonstrates that our methods can generate cohesive hash codes for similar data pairs and dispersed hash codes for dissimilar pairs.

Open vocabulary object detection has been greatly advanced by the recent development of vision-language pretrained model, which helps recognize novel objects with only semantic categories. The prior works mainly focus on knowledge transferring to the object proposal classification and employ class-agnostic box and mask prediction. In this work, we propose CondHead, a principled dynamic network design to better generalize the box regression and mask segmentation for open vocabulary setting. The core idea is to conditionally parameterize the network heads on semantic embedding and thus the model is guided with class-specific knowledge to better detect novel categories. Specifically, CondHead is composed of two streams of network heads, the dynamically aggregated head and dynamically generated head. The former is instantiated with a set of static heads that are conditionally aggregated, these heads are optimized as experts and are expected to learn sophisticated prediction. The latter is instantiated with dynamically generated parameters and encodes general class-specific information. With such a conditional design, the detection model is bridged by the semantic embedding to offer strongly generalizable class-wise box and mask prediction. Our method brings significant improvement to the state-of-the-art open vocabulary object detection methods with very minor overhead, e.g., it surpasses a RegionClip model by 3.0 detection AP on novel categories, with only 1.1% more computation.

Existing self-supervised 3D human pose estimation schemes have largely relied on weak supervisions like consistency loss to guide the learning, which, inevitably, leads to inferior results in real-world scenarios with unseen poses. In this paper, we propose a novel self-supervised approach that allows us to explicitly generate 2D-3D pose pairs for augmenting supervision, through a self-enhancing dual-loop learning framework. This is made possible via introducing a reinforcement-learning-based imitator, which is learned jointly with a pose estimator alongside a pose hallucinator; the three components form two loops during the training process, complementing and strengthening one another. Specifically, the pose estimator transforms an input 2D pose sequence to a low-fidelity 3D output, which is then enhanced by the imitator that enforces physical constraints. The refined 3D poses are subsequently fed to the hallucinator for producing even more diverse data, which are, in turn, strengthened by the imitator and further utilized to train the pose estimator. Such a co-evolution scheme, in practice, enables training a pose estimator on self-generated motion data without relying on any given 3D data. Extensive experiments across various benchmarks demonstrate that our approach yields encouraging results significantly outperforming the state of the art and, in some cases, even on par with results of fully-supervised methods. Notably, it achieves 89.1% 3D PCK on MPI-INF-3DHP under self-supervised cross-dataset evaluation setup, improving upon the previous best self-supervised method [16], [26] by 8.6%.

Although image restoration has achieved significant progress, its potential to assist object detectors in adverse imaging conditions lacks enough attention. It is reported that the existing image restoration methods cannot improve the object detector performance and sometimes even reduce the detection performance. To address the issue, we propose a targeted adversarial attack in the restoration procedure to boost object detection performance after restoration. Specifically, we present an ADAM-like adversarial attack to generate pseudo ground truth for restoration training. Resultant restored images are close to original sharp images, and at the same time, lead to better results of object detection. We conduct extensive experiments in image dehazing and low light enhancement and show the superiority of our method over conventional training and other domain adaptation and multi-task methods. The proposed pipeline can be applied to all restoration methods and detectors in both one- and two-stage.

Recent state-of-the-art one-stage instance segmentation model SOLO divides the input image into a grid and directly predicts per grid cell object masks with fully-convolutional networks, yielding comparably good performance as traditional two-stage Mask R-CNN yet enjoying much simpler architecture and higher efficiency. We observe SOLO generates similar masks for an object at nearby grid cells, and these neighboring predictions can complement each other as some may better segment certain object part, most of which are however directly discarded by non-maximum-suppression. Motivated by the observed gap, we develop a novel learning-based aggregation method that improves upon SOLO by leveraging the rich neighboring information while maintaining the architectural efficiency. The resulting model is named SODAR. Unlike the original per grid cell object masks, SODAR is implicitly supervised to learn mask representations that encode geometric structure of nearby objects and complement adjacent representations with context. The aggregation method further includes two novel designs: 1) a mask interpolation mechanism that enables the model to generate much fewer mask representations by sharing neighboring representations among nearby grid cells, and thus saves computation and memory; 2) a deformable neighbour sampling mechanism that allows the model to adaptively adjust neighbor sampling locations thus gathering mask representations with more relevant context and achieving higher performance. SODAR significantly improves the instance segmentation performance, e.g., it outperforms a SOLO model with ResNet-101 backbone by 2.2 AP on COCO test set, with only about 3% additional computation. We further show consistent performance gain with the SOLOv2 model.

Convolutional neural networks (CNNs) have achieved significant success in the single image dehazing task. Unfortunately, most existing deep dehazing models have high computational complexity, which hinders their application to high-resolution images, especially for UHD (ultra-high-definition) or 4K resolution images. To address the problem, we propose a novel network capable of real-time dehazing of 4K images on a single GPU, which consists of three deep CNNs. The first CNN extracts haze-relevant features at a reduced resolution of the hazy input and then fits locally-affine models in the bilateral space. Another CNN is used to learn multiple full-resolution guidance maps corresponding to the learned bilateral model. As a result, the feature maps with high-frequency can be reconstructed by multi-guided bilateral upsampling. Finally, the third CNN fuses the high-quality feature maps into a dehazed image. In addition, we create a large-scale 4K image dehazing dataset to support the training and testing of compared models. Experimental results demonstrate that the proposed algorithm performs favorably against the state-of-the-art dehazing approaches on various benchmarks.

Transformers, which are popular for language modeling, have been explored for solving vision tasks recently, e.g., the Vision Transformer (ViT) for image classification. The ViT model splits each image into a sequence of tokens with fixed length and then applies multiple Transformer layers to model their global relation for classification. However, ViT achieves inferior performance to CNNs when trained from scratch on a midsize dataset like ImageNet. We find it is because: 1) the simple tokenization of input images fails to model the important local structure such as edges and lines among neighboring pixels, leading to low training sample efficiency; 2) the redundant attention backbone design of ViT leads to limited feature richness for fixed computation budgets and limited training samples. To overcome such limitations, we propose a new Tokens-To-Token Vision Transformer (T2T-VTT), which incorporates 1) a layer-wise Tokens-to-Token (T2T) transformation to progressively structurize the image to tokens by recursively aggregating neighboring Tokens into one Token (Tokens-to-Token), such that local structure represented by surrounding tokens can be modeled and tokens length can be reduced; 2) an efficient backbone with a deep-narrow structure for vision transformer motivated by CNN architecture design after empirical study. Notably, T2T-ViT reduces the parameter count and MACs of vanilla ViT by half, while achieving more than 3.0% improvement when trained from scratch on ImageNet. It also outperforms ResNets and achieves comparable performance with MobileNets by directly training on ImageNet. For example, T2T-ViT with comparable size to ResNet50 (21.5M parameters) can achieve 83.3% top1 accuracy in image resolution 384x384 on ImageNet.1

Recently, DETR [3] pioneered the solution of vision tasks with transformers, it directly translates the image feature map into the object detection result. Though effective, translating the full feature map can be costly due to redundant computation on some area like the background. In this work, we encapsulate the idea of reducing spatial redundancy into a novel poll and pool (PnP) sampling module, with which we build an end-to-end PnP-DETR architecture that adaptively allocates its computation spatially to be more efficient. Concretely, the PnP module abstracts the image feature map into fine foreground object feature vectors and a small number of coarse background contextual feature vectors. The transformer models information interaction within the fine-coarse feature space and translates the features into the detection result. Moreover, the PnP-augmented model can instantly achieve various desired trade-offs between performance and computation with a single model by varying the sampled feature length, without requiring to train multiple models as existing methods. Thus it offers greater flexibility for deployment in diverse scenarios with varying computation constraint. We further validate the generalizability of the PnP module on panoptic segmentation and the recent transformer-based image recognition model ViT [7] and show consistent efficiency gain. We believe our method makes a step for efficient visual analysis with transformers, wherein spatial redundancy is commonly observed. Code and models will be available.

Existing single image high dynamic range (HDR) reconstruction methods attempt to expand the range of illuminance. They are not effective in generating plausible textures and colors in the reconstructed results, especially for high-density pixels in ultra-high-definition (UHD) images. To address these problems, we propose a new HDR reconstruction network for UHD images by collaboratively learning color and texture details. First, we propose a dual-path network to extract the content and chromatic features at a reduced resolution of the low dynamic range (LDR) input. These two types of features are used to fit bilateral-space affine models for real-time HDR reconstruction. To extract the main data structure of the LDR input, we propose to use 3D Tucker decomposition and reconstruction to prevent pseudo edges and noise amplification in the learned bilateral grid. As a result, the high-quality content and chromatic features can be reconstructed capitalized on guided bilateral upsampling. Finally, we fuse these two full-resolution feature maps into the HDR reconstructed results. Our proposed method can achieve real-time processing for UHD images (about 160 fps). Experimental results demonstrate that the proposed algorithm performs favorably against the state-of-the-art HDR reconstruction approaches on public benchmarks and real-world UHD images.

We present Multi-view Pose transformer (MvP) for estimating multi-person 3D poses from multi-view images. Instead of estimating 3D joint locations from costly volumetric representation or reconstructing the per-person 3D pose from multiple detected 2D poses as in previous methods, MvP directly regresses the multi-person 3D poses in a clean and efficient way, without relying on intermediate tasks. Specifically, MvP represents skeleton joints as learnable query embeddings and let them progressively attend to and reason over the multi-view information from the input images to directly regress the actual 3D joint locations. To improve the accuracy of such a simple pipeline, MvP presents a hierarchical scheme to concisely represent query embeddings of multi-person skeleton joints and introduces an input-dependent query adaptation approach. Further, MvP designs a novel geometrically guided attention mechanism, called projective attention, to more precisely fuse the cross-view information for each joint. MvP also introduces a RayConv operation to integrate the view-dependent camera geometry into the feature representations for augmenting the projective attention. We show experimentally that our MvP model outperforms the state-of-the-art methods on several benchmarks while being much more efficient. Notably, it achieves 92.3% AP25 on the challenging Panoptic dataset, improving upon the previous best approach [35] by 9.8%. MvP is general and also extendable to recovering human mesh represented by the SMPL model, thus useful for modeling multi-person body shapes. Code and models are available at https://github.com/sail-sg/mvp.

Solving long-tail large vocabulary object detection with deep learning based models is a challenging and demanding task, which is however under-explored. In this work, we provide the first systematic analysis on the underperformance of state-of-the-art models in front of long-tail distribution. We find existing detection methods are unable to model few-shot classes when the dataset is extremely skewed, which can result in classifier imbalance in terms of parameter magnitude. Directly adapting long-tail classification models to detection frameworks can not solve this problem due to the intrinsic difference between detection and classification. In this work, we propose a novel balanced group softmax (BAGS) module for balancing the classifiers within the detection frameworks through group-wise training. It implicitly modulates the training process for the head and tail classes and ensures they are both sufficiently trained, without requiring any extra sampling for the instances from the tail classes. Extensive experiments on the very recent long-tail large vocabulary object recognition benchmark LVIS show that our proposed BAGS significantly improves the performance of detectors with various backbones and frameworks on both object detection and instance segmentation. It beats all state-of-the-art methods transferred from long-tail image classification and establishes new state-of-the-art. Code is available at https://github.com/FishYuLi/BalancedGroupSoftmax.

Knowledge Distillation (KD) aims to distill the knowledge of a cumbersome teacher model into a lightweight student model. Its success is generally attributed to the privileged information on similarities among categories provided by the teacher model, and in this sense, only strong teacher models are deployed to teach weaker students in practice. In this work, we challenge this common belief by following experimental observations: 1) beyond the acknowledgment that the teacher can improve the student, the student can also enhance the teacher significantly by reversing the KD procedure; 2) a poorly-trained teacher with much lower accuracy than the student can still improve the latter significantly. To explain these observations, we provide a theoretical analysis of the relationships between KD and label smoothing regularization. We prove that 1) KD is a type of learned label smoothing regularization and 2) label smoothing regularization provides a virtual teacher model for KD. From these results, we argue that the success of KD is not fully due to the similarity information between categories from teachers, but also to the regularization of soft targets, which is equally or even more important. Based on these analyses, we further propose a novel Teacher-free Knowledge Distillation (Tf-KD) framework, where a student model learns from itself or manually-designed regularization distribution. The Tf-KD achieves comparable performance with normal KD from a superior teacher, which is well applied when a stronger teacher model is unavailable. Meanwhile, Tf-KD is generic and can be directly deployed for training deep neural networks. Without any extra computation cost, Tf-KD achieves up to 0.65\% improvement on ImageNet over well-established baseline models, which is superior to label smoothing regularization.

Existing data-dependent hashing methods usually learn hash functions from pairwise or triplet data relationships, which only capture the data similarity locally, and often suffer from low learning efficiency and low collision rate. In this work, we propose a new \emph{global} similarity metric, termed as \emph{central similarity}, with which the hash codes of similar data pairs are encouraged to approach a common center and those for dissimilar pairs to converge to different centers, to improve hash learning efficiency and retrieval accuracy. We principally formulate the computation of the proposed central similarity metric by introducing a new concept, i.e., \emph{hash center} that refers to a set of data points scattered in the Hamming space with a sufficient mutual distance between each other. We then provide an efficient method to construct well separated hash centers by leveraging the Hadamard matrix and Bernoulli distributions. Finally, we propose the Central Similarity Quantization (CSQ) that optimizes the central similarity between data points w.r.t.\ their hash centers instead of optimizing the local similarity. CSQ is generic and applicable to both image and video hashing scenarios. Extensive experiments on large-scale image and video retrieval tasks demonstrate that CSQ can generate cohesive hash codes for similar data pairs and dispersed hash codes for dissimilar pairs, achieving a noticeable boost in retrieval performance, i.e. 3\%-20\% in mAP over the previous state-of-the-arts.

Detecting and recognizing human action in videos with crowed scenes is a challenging problem due to the complex environment and diversity events. Prior works always fail to deal with this problem in two aspects: (1) lacking utilizing information of the scenes; (2) lacking training data in the crowd and complex scenes. In this paper, we focus on improving spatio-temporal action recognition by fully-utilizing the information of scenes and collecting new data. A top-down strategy is used to overcome the limitations. Specifically, we adopt a strong human detector to detect the spatial location of each frame. We then apply action recognition models to learn the spatio-temporal information from video frames on both the HIE dataset and new data with diverse scenes from the internet, which can improve the generalization ability of our model. Besides, the scenes information is extracted by the semantic segmentation model to assistant the process. As a result, our method achieved an average 26.05 wf\_mAP (ranking 1st place in the ACM MM grand challenge 2020: Human in Events).

State-of-the-art CNN based recognition models are often computationally prohibitive to deploy on low-end devices. A promising high level approach tackling this limitation is knowledge distillation, which let small student model mimic cumbersome teacher model's output to get improved generalization. However, related methods mainly focus on simple task of classification while do not consider complex tasks like object detection. We show applying the vanilla knowledge distillation to detection model gets minor gain. To address the challenge of distilling knowledge in detection model, we propose a fine-grained feature imitation method exploiting the cross-location discrepancy of feature response. Our intuition is that detectors care more about local near object regions. Thus the discrepancy of feature response on the near object anchor locations reveals important information of how teacher model tends to generalize. We design a novel mechanism to estimate those locations and let student model imitate the teacher on them to get enhanced performance. We first validate the idea on a developed lightweight toy detector which carries simplest notion of current state-of-the-art anchor based detection models on challenging KITTI dataset, our method generates up to 15% boost of mAP for the student model compared to the non-imitated counterpart. We then extensively evaluate the method with Faster R-CNN model under various scenarios with common object detection benchmark of Pascal VOC and COCO, imitation alleviates up to 74% performance drop of student model compared to teacher. Codes released at https://github.com/twangnh/Distilling-Object-Detectors.

To mitigate the detection performance drop caused by domain shift, we aim to develop a novel few-shot adaptation approach that requires only a few target domain images with limited bounding box annotations. To this end, we first observe several significant challenges. First, the target domain data is highly insufficient, making most existing domain adaptation methods ineffective. Second, object detection involves simultaneous localization and classification, further complicating the model adaptation process. Third, the model suffers from over-adaptation (similar to overfitting when training with a few data example) and instability risk that may lead to degraded detection performance in the target domain. To address these challenges, we first introduce a pairing mechanism over source and target features to alleviate the issue of insufficient target domain samples. We then propose a bi-level module to adapt the source trained detector to the target domain: 1) the split pooling based image level adaptation module uniformly extracts and aligns paired local patch features over locations, with different scale and aspect ratio; 2) the instance level adaptation module semantically aligns paired object features while avoids inter-class confusion. Meanwhile, a source model feature regularization (SMFR) is applied to stabilize the adaptation process of the two modules. Combining these contributions gives a novel few-shot adaptive Faster-RCNN framework, termed FAFRCNN, which effectively adapts to target domain with a few labeled samples. Experiments with multiple datasets show that our model achieves new state-of-the-art performance under both the interested few-shot domain adaptation(FDA) and unsupervised domain adaptation(UDA) setting.