Multi-label Learning on Image data has been widely exploited with deep learning models. However, supervised training on deep CNN models often cannot discover sufficient discriminative features for classification. As a result, numerous self-supervision methods are proposed to learn more robust image representations. However, most self-supervised approaches focus on single-instance single-label data and fall short on more complex images with multiple objects. Therefore, we propose an Object-Aware Self-Supervision (OASS) method to obtain more fine-grained representations for multi-label learning, dynamically generating auxiliary tasks based on object locations. Secondly, the robust representation learned by OASS can be leveraged to efficiently generate Class-Specific Instances (CSI) in a proposal-free fashion to better guide multi-label supervision signal transfer to instances. Extensive experiments on the VOC2012 dataset for multi-label classification demonstrate the effectiveness of the proposed method against the state-of-the-art counterparts.
One critical component in lossy deep image compression is the entropy model, which predicts the probability distribution of the quantized latent representation in the encoding and decoding modules. Previous works build entropy models upon convolutional neural networks which are inefficient in capturing global dependencies. In this work, we propose a novel transformer-based entropy model, termed Entroformer, to capture long-range dependencies in probability distribution estimation effectively and efficiently. Different from vision transformers in image classification, the Entroformer is highly optimized for image compression, including a top-k self-attention and a diamond relative position encoding. Meanwhile, we further expand this architecture with a parallel bidirectional context model to speed up the decoding process. The experiments show that the Entroformer achieves state-of-the-art performance on image compression while being time-efficient.
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.
Due to its promising results, density map regression has been widely employed for image-based crowd counting. The approach, however, often suffers from severe performance degradation when tested on data from unseen scenarios, the so-called "domain shift" problem. To address the problem, we investigate in this work single domain generalization (SDG) for crowd counting. The existing SDG approaches are mainly for image classification and segmentation, and can hardly be extended to our case due to its regression nature and label ambiguity (i.e., ambiguous pixel-level ground truths). We propose MPCount, a novel effective SDG approach even for narrow source distribution. MPCount stores diverse density values for density map regression and reconstructs domain-invariant features by means of only one memory bank, a content error mask and attention consistency loss. By partitioning the image into grids, it employs patch-wise classification as an auxiliary task to mitigate label ambiguity. Through extensive experiments on different datasets, MPCount is shown to significantly improve counting accuracy compared to the state of the art under diverse scenarios unobserved in the training data characterized by narrow source distribution. Code is available at https://github.com/Shimmer93/MPCount.
Deep neural networks trained for classification have been found to learn powerful image representations, which are also often used for other tasks such as comparing images w.r.t. their visual similarity. However, visual similarity does not imply semantic similarity. In order to learn semantically discriminative features, we propose to map images onto class embeddings whose pair-wise dot products correspond to a measure of semantic similarity between classes. Such an embedding does not only improve image retrieval results, but could also facilitate integrating semantics for other tasks, e.g., novelty detection or few-shot learning. We introduce a deterministic algorithm for computing the class centroids directly based on prior world-knowledge encoded in a hierarchy of classes such as WordNet. Experiments on CIFAR-100, NABirds, and ImageNet show that our learned semantic image embeddings improve the semantic consistency of image retrieval results by a large margin.
Deep learning has revolutionized computer vision, but it achieved its tremendous success using deep network architectures which are mostly hand-crafted and therefore likely suboptimal. Neural Architecture Search (NAS) aims to bridge this gap by following a well-defined optimization paradigm which systematically looks for the best architecture, given objective criterion such as maximal classification accuracy. The main limitation of NAS is however its astronomical computational cost, as it typically requires training each candidate network architecture from scratch. In this paper, we aim to alleviate this limitation by proposing a novel training-free proxy for image classification accuracy based on Fisher Information. The proposed proxy has a strong theoretical background in statistics and it allows estimating expected image classification accuracy of a given deep network without training the network, thus significantly reducing computational cost of standard NAS algorithms. Our training-free proxy achieves state-of-the-art results on three public datasets and in two search spaces, both when evaluated using previously proposed metrics, as well as using a new metric that we propose which we demonstrate is more informative for practical NAS applications. The source code is publicly available at http://www.github.com/ondratybl/VKDNW
We propose an accurate and fast classification network for classification of brain tumors in MRI images that outperforms all lightweight methods investigated in terms of accuracy. We test our model on a challenging 2D T1-weighted CE-MRI dataset containing three types of brain tumors: Meningioma, Glioma and Pituitary. We introduce an l2-normalized spatial attention mechanism that acts as a regularizer against overfitting during training. We compare our results against the state-of-the-art on this dataset and show that by integrating l2-normalized spatial attention into a baseline network we achieve a performance gain of 1.79 percentage points. Even better accuracy can be attained by combining our model in an ensemble with the pretrained VGG16 at the expense of execution speed. Our code is publicly available at https://github.com/juliadietlmeier/MRI_image_classification
Attention mechanisms, especially self-attention, have played an increasingly important role in deep feature representation for visual tasks. Self-attention updates the feature at each position by computing a weighted sum of features using pair-wise affinities across all positions to capture the long-range dependency within a single sample. However, self-attention has quadratic complexity and ignores potential correlation between different samples. This paper proposes a novel attention mechanism which we call external attention, based on two external, small, learnable, shared memories, which can be implemented easily by simply using two cascaded linear layers and two normalization layers; it conveniently replaces self-attention in existing popular architectures. External attention has linear complexity and implicitly considers the correlations between all data samples. We further incorporate the multi-head mechanism into external attention to provide an all-MLP architecture, external attention MLP (EAMLP), for image classification. Extensive experiments on image classification, object detection, semantic segmentation, instance segmentation, image generation, and point cloud analysis reveal that our method provides results comparable or superior to the self-attention mechanism and some of its variants, with much lower computational and memory costs.
Recent neural architecture search (NAS) based approaches have made great progress in hyperspectral image (HSI) classification tasks. However, the architectures are usually optimized independently of the network weights, increasing searching time and restricting model performances. To tackle these issues, in this paper, different from previous methods that extra define structural parameters, we propose to directly generate structural parameters by utilizing the specifically designed hyper kernels, ingeniously converting the original complex dual optimization problem into easily implemented one-tier optimizations, and greatly shrinking searching costs. Then, we develop a hierarchical multi-module search space whose candidate operations only contain convolutions, and these operations can be integrated into unified kernels. Using the above searching strategy and searching space, we obtain three kinds of networks to separately conduct pixel-level or image-level classifications with 1-D or 3-D convolutions. In addition, by combining the proposed hyper kernel searching scheme with the 3-D convolution decomposition mechanism, we obtain diverse architectures to simulate 3-D convolutions, greatly improving network flexibilities. A series of quantitative and qualitative experiments on six public datasets demonstrate that the proposed methods achieve state-of-the-art results compared with other advanced NAS-based HSI classification approaches.
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.
It is intuitive that NLP tasks for logographic languages like Chinese should benefit from the use of the glyph information in those languages. However, due to the lack of rich pictographic evidence in glyphs and the weak generalization ability of standard computer vision models on character data, an effective way to utilize the glyph information remains to be found. In this paper, we address this gap by presenting Glyce, the glyph-vectors for Chinese character representations. We make three major innovations: (1) We use historical Chinese scripts (e.g., bronzeware script, seal script, traditional Chinese, etc) to enrich the pictographic evidence in characters; (2) We design CNN structures (called tianzege-CNN) tailored to Chinese character image processing; and (3) We use image-classification as an auxiliary task in a multi-task learning setup to increase the model's ability to generalize. We show that glyph-based models are able to consistently outperform word/char ID-based models in a wide range of Chinese NLP tasks. We are able to set new state-of-the-art results for a variety of Chinese NLP tasks, including tagging (NER, CWS, POS), sentence pair classification, single sentence classification tasks, dependency parsing, and semantic role labeling. For example, the proposed model achieves an F1 score of 80.6 on the OntoNotes dataset of NER, +1.5 over BERT; it achieves an almost perfect accuracy of 99.8\% on the Fudan corpus for text classification. Code found at https://github.com/ShannonAI/glyce.
The grade of clear cell renal cell carcinoma (ccRCC) is a critical prognostic factor, making ccRCC nuclei grading a crucial task in RCC pathology analysis. Computer-aided nuclei grading aims to improve pathologists' work efficiency while reducing their misdiagnosis rate by automatically identifying the grades of tumor nuclei within histopathological images. Such a task requires precisely segment and accurately classify the nuclei. However, most of the existing nuclei segmentation and classification methods can not handle the inter-class similarity property of nuclei grading, thus can not be directly applied to the ccRCC grading task. In this paper, we propose a Composite High-Resolution Network for ccRCC nuclei grading. Specifically, we propose a segmentation network called W-Net that can separate the clustered nuclei. Then, we recast the fine-grained classification of nuclei to two cross-category classification tasks, based on two high-resolution feature extractors (HRFEs) which are proposed for learning these two tasks. The two HRFEs share the same backbone encoder with W-Net by a composite connection so that meaningful features for the segmentation task can be inherited for the classification task. Last, a head-fusion block is applied to generate the predicted label of each nucleus. Furthermore, we introduce a dataset for ccRCC nuclei grading, containing 1000 image patches with 70945 annotated nuclei. We demonstrate that our proposed method achieves state-of-the-art performance compared to existing methods on this large ccRCC grading dataset.
Federated learning of deep learning models for supervised tasks, e.g. image classification and segmentation, has found many applications: for example in human-in-the-loop tasks such as film post-production where it enables sharing of domain expertise of human artists in an efficient and effective fashion. In many such applications, we need to protect the training data from being leaked when gradients are shared in the training process due to IP or privacy concerns. Recent works have demonstrated that it is possible to reconstruct the training data from gradients for an image-classification model when its architecture is known. However, there is still an incomplete theoretical understanding of the efficacy and failure of such attacks. In this paper, we analyse the source of training-data leakage from gradients. We formulate the problem of training data reconstruction as solving an optimisation problem iteratively for each layer. The layer-wise objective function is primarily defined by weights and gradients from the current layer as well as the output from the reconstruction of the subsequent layer, but it might also involve a 'pull-back' constraint from the preceding layer. Training data can be reconstructed when we solve the problem backward from the output of the network through each layer. Based on this formulation, we are able to attribute the potential leakage of the training data in a deep network to its architecture. We also propose a metric to measure the level of security of a deep learning model against gradient-based attacks on the training data.
The focus of recent meta-learning research has been on the development of learning algorithms that can quickly adapt to test time tasks with limited data and low computational cost. Few-shot learning is widely used as one of the standard benchmarks in meta-learning. In this work, we show that a simple baseline: learning a supervised or self-supervised representation on the meta-training set, followed by training a linear classifier on top of this representation, outperforms state-of-the-art few-shot learning methods. An additional boost can be achieved through the use of self-distillation. This demonstrates that using a good learned embedding model can be more effective than sophisticated meta-learning algorithms. We believe that our findings motivate a rethinking of few-shot image classification benchmarks and the associated role of meta-learning algorithms. Code is available at: http://github.com/WangYueFt/rfs/.
This research work presents a novel dual-branch model for hyperspectral image classification that combines two streams: one for processing standard hyperspectral patches using Real-Valued Neural Network (RVNN) and the other for processing their corresponding Fourier transforms using Complex-Valued Neural Network (CVNN). The proposed model is evaluated on the Pavia University and Salinas datasets. Results show that the proposed model outperforms state-of-the-art methods in terms of overall accuracy, average accuracy, and Kappa. Through the incorporation of Fourier transforms in the second stream, the model is able to extract frequency information, which complements the spatial information extracted by the first stream. The combination of these two streams improves the overall performance of the model. Furthermore, to enhance the model performance, the Squeeze and Excitation (SE) mechanism has been utilized. Experimental evidence show that SE block improves the models overall accuracy by almost 1\%.
We present lazy visual grounding, a two-stage approach of unsupervised object mask discovery followed by object grounding, for open-vocabulary semantic segmentation. Plenty of the previous art casts this task as pixel-to-text classification without object-level comprehension, leveraging the image-to-text classification capability of pretrained vision-and-language models. We argue that visual objects are distinguishable without the prior text information as segmentation is essentially a vision task. Lazy visual grounding first discovers object masks covering an image with iterative Normalized cuts and then later assigns text on the discovered objects in a late interaction manner. Our model requires no additional training yet shows great performance on five public datasets: Pascal VOC, Pascal Context, COCO-object, COCO-stuff, and ADE 20K. Especially, the visually appealing segmentation results demonstrate the model capability to localize objects precisely. Paper homepage: https://cvlab.postech.ac.kr/research/lazygrounding
Saliency maps that identify the most informative regions of an image for a classifier are valuable for model interpretability. A common approach to creating saliency maps involves generating input masks that mask out portions of an image to maximally deteriorate classification performance, or mask in an image to preserve classification performance. Many variants of this approach have been proposed in the literature, such as counterfactual generation and optimizing over a Gumbel-Softmax distribution. Using a general formulation of masking-based saliency methods, we conduct an extensive evaluation study of a number of recently proposed variants to understand which elements of these methods meaningfully improve performance. Surprisingly, we find that a well-tuned, relatively simple formulation of a masking-based saliency model outperforms many more complex approaches. We find that the most important ingredients for high quality saliency map generation are (1) using both masked-in and masked-out objectives and (2) training the classifier alongside the masking model. Strikingly, we show that a masking model can be trained with as few as 10 examples per class and still generate saliency maps with only a 0.7-point increase in localization error.
Contrastive learning has revolutionized the field of computer vision, learning rich representations from unlabeled data, which generalize well to diverse vision tasks. Consequently, it has become increasingly important to explain these approaches and understand their inner workings mechanisms. Given that contrastive models are trained with interdependent and interacting inputs and aim to learn invariance through data augmentation, the existing methods for explaining single-image systems (e.g., image classification models) are inadequate as they fail to account for these factors and typically assume independent inputs. Additionally, there is a lack of evaluation metrics designed to assess pairs of explanations, and no analytical studies have been conducted to investigate the effectiveness of different techniques used to explaining contrastive learning. In this work, we design visual explanation methods that contribute towards understanding similarity learning tasks from pairs of images. We further adapt existing metrics, used to evaluate visual explanations of image classification systems, to suit pairs of explanations and evaluate our proposed methods with these metrics. Finally, we present a thorough analysis of visual explainability methods for contrastive learning, establish their correlation with downstream tasks and demonstrate the potential of our approaches to investigate their merits and drawbacks.
Histopathology image analysis plays a critical role in cancer diagnosis and treatment. To automatically segment the cancerous regions, fully supervised segmentation algorithms require labor-intensive and time-consuming labeling at the pixel level. In this research, we propose CAMEL, a weakly supervised learning framework for histopathology image segmentation using only image-level labels. Using multiple instance learning (MIL)-based label enrichment, CAMEL splits the image into latticed instances and automatically generates instance-level labels. After label enrichment, the instance-level labels are further assigned to the corresponding pixels, producing the approximate pixel-level labels and making fully supervised training of segmentation models possible. CAMEL achieves comparable performance with the fully supervised approaches in both instance-level classification and pixel-level segmentation on CAMELYON16 and a colorectal adenoma dataset. Moreover, the generality of the automatic labeling methodology may benefit future weakly supervised learning studies for histopathology image analysis.
Glaucoma is one of the most severe eye diseases, characterized by rapid progression and leading to irreversible blindness. It is often the case that diagnostics is carried out when one's sight has already significantly degraded due to the lack of noticeable symptoms at early stage of the disease. Regular glaucoma screenings of the population shall improve early-stage detection, however the desirable frequency of etymological checkups is often not feasible due to the excessive load imposed by manual diagnostics on limited number of specialists. Considering the basic methodology to detect glaucoma is to analyze fundus images for the optic-disc-to-optic-cup ratio, Machine Learning algorithms can offer sophisticated methods for image processing and classification. In our work, we propose an advanced image pre-processing technique combined with a multi-view network of deep classification models to categorize glaucoma. Our Glaucoma Automated Retinal Detection Network (GARDNet) has been successfully tested on Rotterdam EyePACS AIROGS dataset with an AUC of 0.92, and then additionally fine-tuned and tested on RIM-ONE DL dataset with an AUC of 0.9308 outperforming the state-of-the-art of 0.9272. Our code is available on https://github.com/ahmed1996said/gardnet