Face parsing refers to the semantic segmentation of human faces into key facial regions such as eyes, nose, hair, etc. It serves as a prerequisite for various advanced applications, including face editing, face swapping, and facial makeup, which often require segmentation masks for classes like eyeglasses, hats, earrings, and necklaces. These infrequently occurring classes are called long-tail classes, which are overshadowed by more frequently occurring classes known as head classes. Existing methods, primarily CNN-based, tend to be dominated by head classes during training, resulting in suboptimal representation for long-tail classes. Previous works have largely overlooked the problem of poor segmentation performance of long-tail classes. To address this issue, we propose SegFace, a simple and efficient approach that uses a lightweight transformer-based model which utilizes learnable class-specific tokens. The transformer decoder leverages class-specific tokens, allowing each token to focus on its corresponding class, thereby enabling independent modeling of each class. The proposed approach improves the performance of long-tail classes, thereby boosting overall performance. To the best of our knowledge, SegFace is the first work to employ transformer models for face parsing. Moreover, our approach can be adapted for low-compute edge devices, achieving 95.96 FPS. We conduct extensive experiments demonstrating that SegFace significantly outperforms previous state-of-the-art models, achieving a mean F1 score of 88.96 (+2.82) on the CelebAMask-HQ dataset and 93.03 (+0.65) on the LaPa dataset.

Automatic target recognition (ATR) is crucial for safety-critical tasks such as navigation and surveillance, particularly in demanding military scenarios with unfamiliar terrains, harsh environments, and unseen object categories. Conventional and open-world object detectors often fail in these contexts, lacking exposure to such novel conditions. Meanwhile, Large Vision-Language Models (LVLMs) exhibit zero-shot recognition capabilities across diverse settings yet struggle with precise localization. We address these limitations by combining the localization strength of open-world detectors with the recognition confidence of LVLMs, creating a robust pipeline for zero-shot ATR in novel domains and classes. Our study compares several LVLMs on underrepresented military vehicles, examining factors such as distance range, modality, and prompting strategies. These findings provide insights for developing more reliable ATR systems in uncharted environments.

Vision language models (VLMs) respond to user-crafted text prompts and visual inputs, and are applied to numerous real-world problems. VLMs integrate visual modalities with large language models (LLMs), which are well known to be prompt-sensitive. Hence, it is crucial to determine whether VLMs inherit this instability to varying prompts. We therefore investigate which prompt variations VLMs are most sensitive to and which VLMs are most agnostic to prompt variations. To this end, we introduce PARC (Prompt Analysis via Reliability and Calibration), a VLM prompt sensitivity analysis framework built on three pillars: (1) plausible prompt variations in both the language and vision domain, (2) a novel model reliability score with built-in guarantees, and (3) a calibration step that enables dataset-and prompt-spanning prompt variation analysis. Regarding prompt variations, PARC’s evaluation shows that VLMs mirror LLM language prompt sensitivity in the vision domain, and most destructive variations change the expected answer. Regarding models, outstandingly robust VLMs among 22 evaluated models come from the InternVL2 family. We further find indications that prompt sensitivity is linked to training data. https://github.com/NVlabs/PARC

In this work, we introduce FaceXFormer, an end-to-end unified transformer model capable of performing ten facial analysis tasks within a single framework. These tasks include face parsing, landmark detection, head pose estimation, attribute prediction, age, gender, and race estimation, facial expression recognition, face recognition, and face visibility. Traditional face analysis approaches rely on task-specific architectures and pre-processing techniques, limiting scalability and integration. In contrast, FaceXFormer employs a transformer-based encoder-decoder architecture, where each task is represented as a learnable token, enabling seamless multi-task processing within a unified model. To enhance efficiency, we introduce FaceX, a lightweight decoder with a novel bi-directional cross-attention mechanism, which jointly processes face and task tokens to learn robust and generalized facial representations. We train FaceXFormer on ten diverse face perception datasets and evaluate it against both specialized and multi-task models across multiple benchmarks, demonstrating state-of-the-art or competitive performance. Additionally, we analyze the impact of various components of FaceXFormer on performance, assess real-world robustness in "in-the-wild" settings, and conduct a computational performance evaluation. To the best of our knowledge, FaceXFormer is the first model capable of handling ten facial analysis tasks while maintaining real-time performance at 33.21 FPS. Code: https://github.com/Kartik-3004/facexformer

Active Domain Adaptation (ADA) adapts models to target domains by selectively labeling a few target samples. Existing ADA methods prioritize uncertain samples but overlook confident ones, which often match ground-truth. We find that incorporating confident predictions into the labeled set before active sampling reduces the search space and improves adaptation. To address this, we propose a collaborative framework that labels uncertain samples while treating highly confident predictions as ground truth. Our method combines Gaussian Process-based Active Sampling (GPAS) for identifying uncertain samples and Pseudo-Label-based Certain Sampling (PLCS) for confident ones, progressively enhancing adaptation. PLCS refines the search space, and GPAS reduces the domain gap, boosting the proportion of confident samples. Extensive experiments on Office-Home and DomainNet show that our approach outperforms state-of-the-art ADA methods.

Active Learning (AL) aims to enhance the performance of deep models by selecting the most informative samples for annotation from a pool of unlabeled data. Despite impressive performance in closed-set settings, most AL methods fail in real-world scenarios where the unlabeled data contains unknown categories. Recently, a few studies have attempted to tackle the AL problem for the open-set setting. However, these methods focus more on selecting known samples and do not efficiently utilize unknown samples obtained during AL rounds. In this work, we propose an Entropic Open-set AL (EOAL) framework which leverages both known and unknown distributions effectively to select informative samples during AL rounds. Specifically, our approach employs two different entropy scores. One measures the uncertainty of a sample with respect to the known-class distributions. The other measures the uncertainty of the sample with respect to the unknown-class distributions. By utilizing these two entropy scores we effectively separate the known and unknown samples from the unlabeled data resulting in better sampling. Through extensive experiments, we show that the proposed method outperforms existing state-of-the-art methods on CIFAR-10, CIFAR-100, and TinyImageNet datasets. Code is available at https://github.com/bardisafa/EOAL.

Referring Image Segmentation (RIS) aims to segment objects from an image based on a language description. Recent advancements have introduced transformer-based methods that leverage cross-modal dependencies, significantly enhancing performance in referring segmentation tasks. These methods are designed such that each query predicts different masks. However, RIS inherently requires a single-mask prediction, leading to a phenomenon known as Query Collapse, where all queries yield the same mask prediction. This reduces the generalization capability of the RIS model for complex or novel scenarios. To address this issue, we propose a Multi-modal Query Feature Fusion technique, characterized by two innovative designs: (1) Gaussian enhanced Multi-Modal Fusion, a novel visual grounding mechanism that enhances overall representation by extracting rich local visual information and global visual-linguistic relationships, and (2) A Dynamic Query Module that produces a diverse set of queries through a scoring network where the network selectively focuses on queries for objects referred to in the language description. Moreover, we show that including an auxiliary loss to increase the distance between mask representations of different queries further enhances performance and mitigates query collapse. Extensive experiments conducted on four benchmark datasets validate the effectiveness of our framework.

Recent advances in deep learning have led to the development of accurate and efficient models for various computer vision applications such as classification, segmentation, and detection. However, learning highly accurate models relies on the availability of large-scale annotated datasets. Due to this, model performance drops drastically when evaluated on label-scarce datasets having visually distinct images, termed as domain adaptation problem. There are a plethora of works to adapt classification and segmentation models to label-scarce target dataset through unsupervised domain adaptation. Considering that detection is a fundamental task in computer vision, many recent works have focused on developing novel domain adaptive detection techniques. Here, we describe in detail the domain adaptation problem for detection and present an extensive survey of the various methods. Furthermore, we highlight strategies proposed and the associated shortcomings. Subsequently, we identify multiple aspects of the problem that are most promising for future research. We believe that this survey shall be valuable to the pattern recognition experts working in the fields of computer vision, biometrics, medical imaging, and autonomous navigation by introducing them to the problem, and familiarizing them with the current status of the progress while providing promising directions for future research.

Unsupervised Domain Adaptation (UDA) is an effective approach to tackle the issue of domain shift. Specifically, UDA methods try to align the source and target representations to improve generalization on the target domain. Further, UDA methods work under the assumption that the source data is accessible during the adaptation process. However, in real-world scenarios, the labelled source data is often restricted due to privacy regulations, data transmission constraints, or proprietary data concerns. The Source-Free Domain Adaptation (SFDA) setting aims to alleviate these concerns by adapting a source-trained model for the target domain without requiring access to the source data. In this paper, we explore the SFDA setting for the task of adaptive object detection. To this end, we propose a novel training strategy for adapting a source-trained object detector to the target domain without source data. More precisely, we design a novel contrastive loss to enhance the target representations by exploiting the objects relations for a given target domain input. These object instance relations are modelled using an Instance Relation Graph (IRG) network, which are then used to guide the contrastive representation learning. In addition, we utilize a student-teacher to effectively distill knowledge from source-trained model to target domain. Extensive experiments on multiple object detection benchmark datasets show that the proposed approach is able to efficiently adapt source-trained object detectors to the target domain, outperforming state-of-the-art domain adaptive detection methods. Code and models are provided in https://viudomain.github.io/irg-sfda-web/.

Existing instance segmentation models learn task-specific information using manual mask annotations from base (training) categories. These mask annotations require tremendous human effort, limiting the scalability to annotate novel (new) categories. To alleviate this problem, Open-Vocabulary (OV) methods leverage large-scale image-caption pairs and vision-language models to learn novel categories. In summary, an OV method learns task-specific information using strong supervision from base annotations and novel category information using weak supervision from image-captions pairs. This difference between strong and weak supervision leads to overfitting on base categories, resulting in poor generalization towards novel categories. In this work, we overcome this issue by learning both base and novel categories from pseudo-mask annotations generated by the vision-language model in a weakly supervised manner using our proposed Mask-free OVIS pipeline. Our method automatically generates pseudo-mask annotations by leveraging the localization ability of a pre-trained vision-language model for objects present in image-caption pairs. The generated pseudo-mask annotations are then used to supervise an instance segmentation model, freeing the entire pipeline from any labour-expensive instance-level annotations and overfitting. Our extensive experiments show that our method trained with just pseudo-masks significantly improves the mAP scores on the MS-COCO dataset and OpenImages dataset compared to the recent state-of-the-art methods trained with manual masks. Codes and models are provided in https://vibashan.github.io/ovis-web/.

Automatic Target Recognition (ATR) is a category of computer vision algorithms which attempts to recognize targets on data obtained from different sensors. ATR algorithms are extensively used in real-world scenarios such as military and surveillance applications. Existing ATR algorithms are developed for traditional closed-set methods where training and testing have the same class distribution. Thus, these algorithms have not been robust to unknown classes not seen during the training phase, limiting their utility in real-world applications. To this end, we propose an Open-set Automatic Target Recognition framework where we enable open-set recognition capability for ATR algorithms. In addition, we introduce a plugin Category-aware Binary Classifier (CBC) module to effectively tackle unknown classes seen during inference. The proposed CBC module can be easily integrated with any existing ATR algorithms and can be trained in an end-to-end manner. Experimental results show that the proposed approach outperforms many open-set methods on the DSIAC and CIFAR-10 datasets. To the best of our knowledge, this is the first work to address the open-set classification problem for ATR algorithms. Source code is available at: https://github.com/bardisafa/Open-set-ATR.

Existing object detection models assume both the training and test data are sampled from the same source do-main. This assumption does not hold true when these detectors are deployed in real-world applications, where they en-counter new visual domains. Unsupervised Domain Adaptation (UDA) methods are generally employed to mitigate the adverse effects caused by domain shift. Existing UDA methods operate in an offline manner where the model is first adapted toward the target domain and then deployed in real-world applications. However, this offline adaptation strategy is not suitable for real-world applications as the model frequently encounters new domain shifts. Hence, it is critical to develop a feasible UDA method that generalizes to the new domain shifts encountered during deployment time in a continuous online manner. To this end, we propose a novel unified adaptation framework that adapts and improves generalization on the target domain in both offline and online settings. Specifically, we introduce MemXformer - a cross-attention transformer-based memory module where items in the memory take advantage of domain shifts and record prototypical patterns of the target distribution. Further, MemXformer produces strong positive and negative pairs to guide a novel contrastive loss, which enhances target-specific representation learning. Experiments on diverse detection benchmarks show that the proposed strategy producs state-of-the-art performance in both offline and online settings. To the best of our knowledge, this is the first work to address online and offline adaptation settings for object detection. Source code: https://github.com/Vibashan/memXformer-online-da

Unsupervised domain adaptive object detection methods transfer knowledge from the labelled source domain to a visually distinct and unlabeled target domain. Most methods achieve this by training the detector model with the help of both labeled source and unlabeled target data. However, in real-world scenarios, gaining access to source data is not practical due to privacy concerns, legal issues and inefficient data transmission. To this end, we tackle the problem of Source-Free Domain Adaptive Object Detection, where during adaptation, we do not have access to the source data but only the source trained model. Specifically, we introduce Mixture of Teacher Experts (MoTE) method, where our key idea is to exploit the prediction uncertainty through a mixture of teacher models and progressively train the student model. We evaluate the proposed method by conducting extensive experiments on several object detection benchmark datasets to demonstrate the effectiveness of the proposed mixture of teacher expert based student-teacher training, specifically for source-free adaptation.

In image fusion, images obtained from different sensors are fused to generate a single image with enhanced information. In recent years, state-of-the-art methods have adopted Convolution Neural Networks (CNNs) to encode meaningful features for image fusion. Specifically, CNN-based methods perform image fusion by fusing local features. However, they do not consider long-range dependencies that are present in the image. Transformer-based models are designed to overcome this by modelling the long-range dependencies with the help of self-attention mechanism. This motivates us to propose a novel Image Fusion Transformer (IFT) where we develop a transformer-based multi-scale fusion strategy that attends to both local and long-range information (or global context). The proposed method follows a two-stage training approach. In the first stage, we train an auto-encoder to extract deep features at multiple scales. In the second stage, multi-scale features are fused using a Spatio-Transformer (ST) fusion strategy. The ST fusion blocks are comprised of a CNN and a transformer branch which captures local and long-range features, respectively. Extensive experiments on multiple benchmark datasets show that the proposed method performs better than many competitive fusion algorithms. Furthermore, we show the effectiveness of the proposed ST fusion strategy with an ablation analysis.1

Object detectors trained on large-scale RGB datasets are being extensively employed in real-world applications. However, these RGB-trained models suffer a performance drop under adverse illumination and lighting conditions. Infrared (IR) cameras are robust under such conditions and can be helpful in real-world applications. Though thermal cameras are widely used for military applications and increasingly for commercial applications, there is a lack of robust algorithms to robustly exploit the thermal imagery due to the limited availability of labeled thermal data. In this work, we aim to enhance the object detection performance in the thermal domain by leveraging the labeled visible domain data in an Unsupervised Domain Adaptation (UDA) setting. We propose an algorithm agnostic meta-learning framework to improve existing UDA methods instead of proposing a new UDA strategy. We achieve this by meta-learning the initial condition of the detector, which facilitates the adaptation process with fine updates without overfitting or getting stuck at local optima. However, meta-learning the initial condition for the detection scenario is computationally heavy due to long and intractable computation graphs. Therefore, we propose an online meta-learning paradigm which performs online updates resulting in a short and tractable computation graph. To this end, we demonstrate the superiority of our method over many baselines in the UDA setting, producing a state-of-the-art thermal detector for the KAIST and DSIAC datasets.

Existing approaches for unsupervised domain adaptive object detection perform feature alignment via adversarial training. While these methods achieve reasonable improvements in performance, they typically perform category-agnostic domain alignment, thereby resulting in negative transfer of features. To overcome this issue, in this work, we attempt to incorporate category information into the domain adaptation process by proposing Memory Guided Attention for Category-Aware Domain Adaptation (MeGA-CDA). The proposed method consists of employing category-wise discriminators to ensure category-aware feature alignment for learning domain-invariant discriminative features. However, since the category information is not available for the target samples, we propose to generate memory-guided category-specific attention maps which are then used to route the features appropriately to the corresponding category discriminator. The proposed method is evaluated on several benchmark datasets and is shown to outperform existing approaches.

Surgical scene understanding and multi-tasking learning are crucial for image-guided robotic surgery. Training a real-time robotic system for the detection and segmentation of high-resolution images provides a challenging problem with the limited computational resource. The perception drawn can be applied in effective real-time feedback, surgical skill assessment, and human-robot collaborative surgeries to enhance surgical outcomes. For this purpose, we develop a novel end-to-end trainable real-time Multi-Task Learning (MTL) model with weight-shared encoder and task-aware detection and segmentation decoders. Optimization of multiple tasks at the same convergence point is vital and presents a complex problem. Thus, we propose an asynchronous task-aware optimization (ATO) technique to calculate task-oriented gradients and train the decoders independently. Moreover, MTL models are always computationally expensive, which hinder real-time applications. To address this challenge, we introduce a global attention dynamic pruning (GADP) by removing less significant and sparse parameters. We further design a skip squeeze and excitation (SE) module, which suppresses weak features, excites significant features and performs dynamic spatial and channel-wise feature re-calibration. Validating on the robotic instrument segmentation dataset of MICCAI endoscopic vision challenge, our model significantly outperforms state-of-the-art segmentation and detection models, including best-performed models in the challenge.