Low-resolution fine-grained image classification has recently made significant progress, largely thanks to the super-resolution techniques and knowledge distillation methods. However, these approaches lead to an exponential increase in the number of parameters and computational complexity of models. In order to solve this problem, in this letter, we propose a Vision Mamba Distillation (ViMD) approach to enhance the effectiveness and efficiency of low-resolution fine-grained image classification. Concretely, a lightweight super-resolution vision Mamba classification network (SRVM-Net) is proposed to improve its capability for extracting visual features by redesigning the classification sub-network with Mamba modeling. Moreover, we design a novel multi-level Mamba knowledge distillation loss boosting the performance, which can transfer prior knowledge obtained from a High-resolution Vision Mamba classification Network (HRVM-Net) as a teacher into the proposed SRVM-Net as a student. Extensive experiments on seven public fine-grained classification datasets related to benchmarks confirm our ViMD achieves a new state-of-the-art performance. While having higher accuracy, ViMD outperforms similar methods with fewer parameters and FLOPs, which is more suitable for embedded device applications. Code is available at https://github.com/boa2004plaust/ViMD.
Convolutional neural networks (CNNs) are extremely popular and effective for image classification tasks but tend to be overly confident in their predictions. Various works have sought to quantify uncertainty associated with these models, detect out-of-distribution (OOD) inputs, or identify anomalous regions in an image, but limited work has sought to develop a holistic approach that can accurately estimate perception model confidence across various sources of uncertainty. We develop a probabilistic and reconstruction-based competency estimation (PaRCE) method and compare it to existing approaches for uncertainty quantification and OOD detection. We find that our method can best distinguish between correctly classified, misclassified, and OOD samples with anomalous regions, as well as between samples with visual image modifications resulting in high, medium, and low prediction accuracy. We describe how to extend our approach for anomaly localization tasks and demonstrate the ability of our approach to distinguish between regions in an image that are familiar to the perception model from those that are unfamiliar. We find that our method generates interpretable scores that most reliably capture a holistic notion of perception model confidence.
Mamba has shown great potential for computer vision due to its linear complexity in modeling the global context with respect to the input length. However, existing lightweight Mamba-based backbones cannot demonstrate performance that matches Convolution or Transformer-based methods. We observe that simply modifying the scanning path in the image domain is not conducive to fully exploiting the potential of vision Mamba. In this paper, we first perform comprehensive spectral and quantitative analyses, and verify that the Mamba block mainly models low-frequency information under Convolution-Mamba hybrid architecture. Based on the analyses, we introduce a novel Laplace mixer to decouple the features in terms of frequency and input only the low-frequency components into the Mamba block. In addition, considering the redundancy of the features and the different requirements for high-frequency details and low-frequency global information at different stages, we introduce a frequency ramp inception, i.e., gradually reduce the input dimensions of the high-frequency branches, so as to efficiently trade-off the high-frequency and low-frequency components at different layers. By integrating mobile-friendly convolution and efficient Laplace mixer, we build a series of tiny hybrid vision Mamba called TinyViM. The proposed TinyViM achieves impressive performance on several downstream tasks including image classification, semantic segmentation, object detection and instance segmentation. In particular, TinyViM outperforms Convolution, Transformer and Mamba-based models with similar scales, and the throughput is about 2-3 times higher than that of other Mamba-based models. Code is available at https://github.com/xwmaxwma/TinyViM.
Recent advances in vision-language models (VLM) have demonstrated remarkable capability in image classification. These VLMs leverage a predefined set of categories to construct text prompts for zero-shot reasoning. However, in more open-ended domains like autonomous driving, using a predefined set of labels becomes impractical, as the semantic label space is unknown and constantly evolving. Additionally, fixed embedding text prompts often tend to predict a single label (while in reality, multiple labels commonly exist per image). In this paper, we introduce CoA, an innovative Chain-of-Action (CoA) method that generates labels aligned with all contextually relevant features of an image. CoA is designed based on the observation that enriched and valuable contextual information improves generative performance during inference. Traditional vision-language models tend to output singular and redundant responses. Therefore, we employ a tailored CoA to alleviate this problem. We first break down the generative labeling task into detailed actions and construct an CoA leading to the final generative objective. Each action extracts and merges key information from the previous action and passes the enriched information as context to the next action, ultimately improving the VLM in generating comprehensive and accurate semantic labels. We assess the effectiveness of CoA through comprehensive evaluations on widely-used benchmark datasets and the results demonstrate significant improvements across key performance metrics.
Machine learning models are increasingly being utilized across various fields and tasks due to their outstanding performance and strong generalization capabilities. Nonetheless, their success hinges on the availability of large volumes of annotated data, the creation of which is often labor-intensive, time-consuming, and expensive. Many active learning (AL) approaches have been proposed to address these challenges, but they often fail to fully leverage the information from the core phases of AL, such as training on the labeled set and querying new unlabeled samples. To bridge this gap, we propose a novel AL approach, Loss Prediction Loss with Gradient Norm (LPLgrad), designed to quantify model uncertainty effectively and improve the accuracy of image classification tasks. LPLgrad operates in two distinct phases: (i) {\em Training Phase} aims to predict the loss for input features by jointly training a main model and an auxiliary model. Both models are trained on the labeled data to maximize the efficiency of the learning process, an aspect often overlooked in previous AL methods. This dual-model approach enhances the ability to extract complex input features and learn intrinsic patterns from the data effectively; (ii) {\em Querying Phase} that quantifies the uncertainty of the main model to guide sample selection. This is achieved by calculating the gradient norm of the entropy values for samples in the unlabeled dataset. Samples with the highest gradient norms are prioritized for labeling and subsequently added to the labeled set, improving the model's performance with minimal labeling effort. Extensive evaluations on real-world datasets demonstrate that the LPLgrad approach outperforms state-of-the-art methods by order of magnitude in terms of accuracy on a small number of labeled images, yet achieving comparable training and querying times in multiple image classification tasks.
Recent advancements in vision-language models (VLMs), such as CLIP, have demonstrated substantial success in self-supervised representation learning for vision tasks. However, effectively adapting VLMs to downstream applications remains challenging, as their accuracy often depends on time-intensive and expertise-demanding prompt engineering, while full model fine-tuning is costly. This is particularly true for biomedical images, which, unlike natural images, typically suffer from limited annotated datasets, unintuitive image contrasts, and nuanced visual features. Recent prompt learning techniques, such as Context Optimization (CoOp) intend to tackle these issues, but still fall short in generalizability. Meanwhile, explorations in prompt learning for biomedical image analysis are still highly limited. In this work, we propose BiomedCoOp, a novel prompt learning framework that enables efficient adaptation of BiomedCLIP for accurate and highly generalizable few-shot biomedical image classification. Our approach achieves effective prompt context learning by leveraging semantic consistency with average prompt ensembles from Large Language Models (LLMs) and knowledge distillation with a statistics-based prompt selection strategy. We conducted comprehensive validation of our proposed framework on 11 medical datasets across 9 modalities and 10 organs against existing state-of-the-art methods, demonstrating significant improvements in both accuracy and generalizability. The code will be publicly available at https://github.com/HealthX-Lab/BiomedCoOp.
Machine Unlearning (MU) aims to selectively erase harmful behaviors from models while retaining the overall utility of the model. As a multi-task learning problem, MU involves balancing objectives related to forgetting specific concepts/data and preserving general performance. A naive integration of these forgetting and preserving objectives can lead to gradient conflicts, impeding MU algorithms from reaching optimal solutions. To address the gradient conflict issue, we reformulate MU as a two-player cooperative game, where the two players, namely, the forgetting player and the preservation player, contribute via their gradient proposals to maximize their overall gain. To this end, inspired by the Nash bargaining theory, we derive a closed-form solution to guide the model toward the Pareto front, effectively avoiding the gradient conflicts. Our formulation of MU guarantees an equilibrium solution, where any deviation from the final state would lead to a reduction in the overall objectives for both players, ensuring optimality in each objective. We evaluate our algorithm's effectiveness on a diverse set of tasks across image classification and image generation. Extensive experiments with ResNet, vision-language model CLIP, and text-to-image diffusion models demonstrate that our method outperforms state-of-the-art MU algorithms, achieving superior performance on several benchmarks. For example, in the challenging scenario of sample-wise forgetting, our algorithm approaches the gold standard retrain baseline. Our results also highlight improvements in forgetting precision, preservation of generalization, and robustness against adversarial attacks.
As a prominent direction of Artificial General Intelligence (AGI), Multimodal Large Language Models (MLLMs) have garnered increased attention from both industry and academia. Building upon pre-trained LLMs, this family of models further develops multimodal perception and reasoning capabilities that are impressive, such as writing code given a flow chart or creating stories based on an image. In the development process, evaluation is critical since it provides intuitive feedback and guidance on improving models. Distinct from the traditional train-eval-test paradigm that only favors a single task like image classification, the versatility of MLLMs has spurred the rise of various new benchmarks and evaluation methods. In this paper, we aim to present a comprehensive survey of MLLM evaluation, discussing four key aspects: 1) the summarised benchmarks types divided by the evaluation capabilities, including foundation capabilities, model self-analysis, and extented applications; 2) the typical process of benchmark counstruction, consisting of data collection, annotation, and precautions; 3) the systematic evaluation manner composed of judge, metric, and toolkit; 4) the outlook for the next benchmark. This work aims to offer researchers an easy grasp of how to effectively evaluate MLLMs according to different needs and to inspire better evaluation methods, thereby driving the progress of MLLM research.
Few-shot learning presents a critical solution for cancer diagnosis in computational pathology (CPath), addressing fundamental limitations in data availability, particularly the scarcity of expert annotations and patient privacy constraints. A key challenge in this paradigm stems from the inherent disparity between the limited training set of whole slide images (WSIs) and the enormous number of contained patches, where a significant portion of these patches lacks diagnostically relevant information, potentially diluting the model's ability to learn and focus on critical diagnostic features. While recent works attempt to address this by incorporating additional knowledge, several crucial gaps hinder further progress: (1) despite the emergence of powerful pathology foundation models (FMs), their potential remains largely untapped, with most approaches limiting their use to basic feature extraction; (2) current language guidance mechanisms attempt to align text prompts with vast numbers of WSI patches all at once, struggling to leverage rich pathological semantic information. To this end, we introduce the knowledge-enhanced adaptive visual compression framework, dubbed FOCUS, which uniquely combines pathology FMs with language prior knowledge to enable a focused analysis of diagnostically relevant regions by prioritizing discriminative WSI patches. Our approach implements a progressive three-stage compression strategy: we first leverage FMs for global visual redundancy elimination, and integrate compressed features with language prompts for semantic relevance assessment, then perform neighbor-aware visual token filtering while preserving spatial coherence. Extensive experiments on pathological datasets spanning breast, lung, and ovarian cancers demonstrate its superior performance in few-shot pathology diagnosis. Codes are available at https://github.com/dddavid4real/FOCUS.
Geometric shape features have been widely used as strong predictors for image classification. Nevertheless, most existing classifiers such as deep neural networks (DNNs) directly leverage the statistical correlations between these shape features and target variables. However, these correlations can often be spurious and unstable across different environments (e.g., in different age groups, certain types of brain changes have unstable relations with neurodegenerative disease); hence leading to biased or inaccurate predictions. In this paper, we introduce a novel framework that for the first time develops invariant shape representation learning (ISRL) to further strengthen the robustness of image classifiers. In contrast to existing approaches that mainly derive features in the image space, our model ISRL is designed to jointly capture invariant features in latent shape spaces parameterized by deformable transformations. To achieve this goal, we develop a new learning paradigm based on invariant risk minimization (IRM) to learn invariant representations of image and shape features across multiple training distributions/environments. By embedding the features that are invariant with regard to target variables in different environments, our model consistently offers more accurate predictions. We validate our method by performing classification tasks on both simulated 2D images, real 3D brain and cine cardiovascular magnetic resonance images (MRIs). Our code is publicly available at https://github.com/tonmoy-hossain/ISRL.
Various linear complexity models, such as Linear Transformer (LinFormer), State Space Model (SSM), and Linear RNN (LinRNN), have been proposed to replace the conventional softmax attention in Transformer structures. However, the optimal design of these linear models is still an open question. In this work, we attempt to answer this question by finding the best linear approximation to softmax attention from a theoretical perspective. We start by unifying existing linear complexity models as the linear attention form and then identify three conditions for the optimal linear attention design: 1) Dynamic memory ability; 2) Static approximation ability; 3) Least parameter approximation. We find that none of the current linear models meet all three conditions, resulting in suboptimal performance. Instead, we propose Meta Linear Attention (MetaLA) as a solution that satisfies these conditions. Our experiments on Multi-Query Associative Recall (MQAR) task, language modeling, image classification, and Long-Range Arena (LRA) benchmark demonstrate that MetaLA is more effective than the existing linear models.
In computer vision tasks, the ability to focus on relevant regions within an image is crucial for improving model performance, particularly when key features are small, subtle, or spatially dispersed. Convolutional neural networks (CNNs) typically treat all regions of an image equally, which can lead to inefficient feature extraction. To address this challenge, I have introduced Vision Eagle Attention, a novel attention mechanism that enhances visual feature extraction using convolutional spatial attention. The model applies convolution to capture local spatial features and generates an attention map that selectively emphasizes the most informative regions of the image. This attention mechanism enables the model to focus on discriminative features while suppressing irrelevant background information. I have integrated Vision Eagle Attention into a lightweight ResNet-18 architecture, demonstrating that this combination results in an efficient and powerful model. I have evaluated the performance of the proposed model on three widely used benchmark datasets: FashionMNIST, Intel Image Classification, and OracleMNIST, with a primary focus on image classification. Experimental results show that the proposed approach improves classification accuracy. Additionally, this method has the potential to be extended to other vision tasks, such as object detection, segmentation, and visual tracking, offering a computationally efficient solution for a wide range of vision-based applications. Code is available at: https://github.com/MahmudulHasan11085/Vision-Eagle-Attention.git
Given sufficient data from multiple edge devices, federated learning (FL) enables training a shared model without transmitting private data to a central server. However, FL is generally vulnerable to Byzantine attacks from compromised edge devices, which can significantly degrade the model performance. In this paper, we propose a intuitive plugin that can be integrated into existing FL techniques to achieve Byzantine-Resilience. Key idea is to generate virtual data samples and evaluate model consistency scores across local updates to effectively filter out compromised edge devices. By utilizing this scoring mechanism before the aggregation phase, the proposed plugin enables existing FL techniques to become robust against Byzantine attacks while maintaining their original benefits. Numerical results on medical image classification task validate that plugging the proposed approach into representative FL algorithms, effectively achieves Byzantine resilience. Furthermore, the proposed plugin maintains the original convergence properties of the base FL algorithms when no Byzantine attacks are present.
Vision Transformers (ViTs) have redefined image classification by leveraging self-attention to capture complex patterns and long-range dependencies between image patches. However, a key challenge for ViTs is efficiently incorporating multi-scale feature representations, which is inherent in convolutional neural networks (CNNs) through their hierarchical structure. Graph transformers have made strides in addressing this by leveraging graph-based modeling, but they often lose or insufficiently represent spatial hierarchies, especially since redundant or less relevant areas dilute the image's contextual representation. To bridge this gap, we propose SAG-ViT, a Scale-Aware Graph Attention ViT that integrates multi-scale feature capabilities of CNNs, representational power of ViTs, graph-attended patching to enable richer contextual representation. Using EfficientNetV2 as a backbone, the model extracts multi-scale feature maps, dividing them into patches to preserve richer semantic information compared to directly patching the input images. The patches are structured into a graph using spatial and feature similarities, where a Graph Attention Network (GAT) refines the node embeddings. This refined graph representation is then processed by a Transformer encoder, capturing long-range dependencies and complex interactions. We evaluate SAG-ViT on benchmark datasets across various domains, validating its effectiveness in advancing image classification tasks. Our code and weights are available at https://github.com/shravan-18/SAG-ViT.
Vision Transformers (ViT) have recently brought a new wave of research in the field of computer vision. These models have done particularly well in the field of image classification and segmentation. Research on semantic and instance segmentation has emerged to accelerate with the inception of the new architecture, with over 80\% of the top 20 benchmarks for the iSAID dataset being either based on the ViT architecture or the attention mechanism behind its success. This paper focuses on the heuristic comparison of three key factors of using (or not using) ViT for semantic segmentation of remote sensing aerial images on the iSAID. The experimental results observed during the course of the research were under the scrutinization of the following objectives: 1. Use of weighted fused loss function for the maximum mean Intersection over Union (mIoU) score, Dice score, and minimization or conservation of entropy or class representation, 2. Comparison of transfer learning on Meta's MaskFormer, a ViT-based semantic segmentation model, against generic UNet Convolutional Neural Networks (CNNs) judged over mIoU, Dice scores, training efficiency, and inference time, and 3. What do we lose for what we gain? i.e., the comparison of the two models against current state-of-art segmentation models. We show the use of the novel combined weighted loss function significantly boosts the CNN model's performance capacities as compared to transfer learning the ViT. The code for this implementation can be found on \url{https://github.com/ashimdahal/ViT-vs-CNN-ImageSegmentation}.
Graph Neural Networks (GNNs) have advanced relational data analysis but lack invariance learning techniques common in image classification. In node classification with GNNs, it is actually the ego-graph of the center node that is classified. This research extends the scale invariance concept to node classification by drawing an analogy to image processing: just as scale invariance being used in image classification to capture multi-scale features, we propose the concept of ``scaled ego-graphs''. Scaled ego-graphs generalize traditional ego-graphs by replacing undirected single-edges with ``scaled-edges'', which are ordered sequences of multiple directed edges. We empirically assess the performance of the proposed scale invariance in graphs on seven benchmark datasets, across both homophilic and heterophilic structures. Our scale-invariance-based graph learning outperforms inception models derived from random walks by being simpler, faster, and more accurate. The scale invariance explains inception models' success on homophilic graphs and limitations on heterophilic graphs. To ensure applicability of inception model to heterophilic graphs as well, we further present ScaleNet, an architecture that leverages multi-scaled features. ScaleNet achieves state-of-the-art results on five out of seven datasets (four homophilic and one heterophilic) and matches top performance on the remaining two, demonstrating its excellent applicability. This represents a significant advance in graph learning, offering a unified framework that enhances node classification across various graph types. Our code is available at https://github.com/Qin87/ScaleNet/tree/July25.
Fine-grained classification of whole slide images (WSIs) is essential in precision oncology, enabling precise cancer diagnosis and personalized treatment strategies. The core of this task involves distinguishing subtle morphological variations within the same broad category of gigapixel-resolution images, which presents a significant challenge. While the multi-instance learning (MIL) paradigm alleviates the computational burden of WSIs, existing MIL methods often overlook hierarchical label correlations, treating fine-grained classification as a flat multi-class classification task. To overcome these limitations, we introduce a novel hierarchical multi-instance learning (HMIL) framework. By facilitating on the hierarchical alignment of inherent relationships between different hierarchy of labels at instance and bag level, our approach provides a more structured and informative learning process. Specifically, HMIL incorporates a class-wise attention mechanism that aligns hierarchical information at both the instance and bag levels. Furthermore, we introduce supervised contrastive learning to enhance the discriminative capability for fine-grained classification and a curriculum-based dynamic weighting module to adaptively balance the hierarchical feature during training. Extensive experiments on our large-scale cytology cervical cancer (CCC) dataset and two public histology datasets, BRACS and PANDA, demonstrate the state-of-the-art class-wise and overall performance of our HMIL framework. Our source code is available at https://github.com/ChengJin-git/HMIL.
In this paper, we question if well pre-trained vision transformer (ViT) models could be used as teachers that exhibit scalable properties to advance cross architecture knowledge distillation (KD) research, in the context of using large-scale datasets for evaluation. To make this possible, our analysis underlines the importance of seeking effective strategies to align (1) feature computing paradigm differences, (2) model scale differences, and (3) knowledge density differences. By combining three coupled components namely cross attention projector, dual-view feature mimicking and teacher parameter perception tailored to address the above problems, we present a simple and effective KD method, called ScaleKD. Our method can train student backbones that span across a variety of convolutional neural network (CNN), multi-layer perceptron (MLP), and ViT architectures on image classification datasets, achieving state-of-the-art distillation performance. For instance, taking a well pre-trained Swin-L as the teacher model, our method gets 75.15%|82.03%|84.16%|78.63%|81.96%|83.93%|83.80%|85.53% top-1 accuracies for MobileNet-V1|ResNet-50|ConvNeXt-T|Mixer-S/16|Mixer-B/16|ViT-S/16|Swin-T|ViT-B/16 models trained on ImageNet-1K dataset from scratch, showing 3.05%|3.39%|2.02%|4.61%|5.52%|4.03%|2.62%|3.73% absolute gains to the individually trained counterparts. Intriguingly, when scaling up the size of teacher models or their pre-training datasets, our method showcases the desired scalable properties, bringing increasingly larger gains to student models. The student backbones trained by our method transfer well on downstream MS-COCO and ADE20K datasets. More importantly, our method could be used as a more efficient alternative to the time-intensive pre-training paradigm for any target student model if a strong pre-trained ViT is available, reducing the amount of viewed training samples up to 195x.
Deep learning (DL) models are popular across various domains due to their remarkable performance and efficiency. However, their effectiveness relies heavily on large amounts of labeled data, which are often time-consuming and labor-intensive to generate manually. To overcome this challenge, it is essential to develop strategies that reduce reliance on extensive labeled data while preserving model performance. In this paper, we propose FisherMask, a Fisher information-based active learning (AL) approach that identifies key network parameters by masking them based on their Fisher information values. FisherMask enhances batch AL by using Fisher information to select the most critical parameters, allowing the identification of the most impactful samples during AL training. Moreover, Fisher information possesses favorable statistical properties, offering valuable insights into model behavior and providing a better understanding of the performance characteristics within the AL pipeline. Our extensive experiments demonstrate that FisherMask significantly outperforms state-of-the-art methods on diverse datasets, including CIFAR-10 and FashionMNIST, especially under imbalanced settings. These improvements lead to substantial gains in labeling efficiency. Hence serving as an effective tool to measure the sensitivity of model parameters to data samples. Our code is available on \url{https://github.com/sgchr273/FisherMask}.
Convolutional Neural Networks (CNNs) have seen significant performance improvements in recent years. However, due to their size and complexity, they function as black-boxes, leading to transparency concerns. State-of-the-art saliency methods generate local explanations that highlight the area in the input image where a class is identified but cannot explain how a concept of interest contributes to the prediction, which is essential for bias mitigation. On the other hand, concept-based methods, such as TCAV (Testing with Concept Activation Vectors), provide insights into how sensitive is the network to a concept, but cannot compute its attribution in a specific prediction nor show its location within the input image. This paper introduces a novel post-hoc explainability framework, Visual-TCAV, which aims to bridge the gap between these methods by providing both local and global explanations for CNN-based image classification. Visual-TCAV uses Concept Activation Vectors (CAVs) to generate saliency maps that show where concepts are recognized by the network. Moreover, it can estimate the attribution of these concepts to the output of any class using a generalization of Integrated Gradients. This framework is evaluated on popular CNN architectures, with its validity further confirmed via experiments where ground truth for explanations is known, and a comparison with TCAV. Our code will be made available soon.