Fine-tuning vision-language models (VLMs) with large amounts of unlabeled data has recently garnered significant interest. However, a key challenge remains the lack of high-quality pseudo-labeled data. Current pseudo-labeling strategies often struggle with mismatches between semantic and visual information, leading to sub-optimal performance of unsupervised prompt learning (UPL) methods. In this paper, we introduce a simple yet effective approach called \textbf{A}ugmenting D\textbf{i}scriminative \textbf{R}ichness via Diffusions (AiR), toward learning a richer discriminating way to represent the class comprehensively and thus facilitate classification. Specifically, our approach includes a pseudo-label generation module that leverages high-fidelity synthetic samples to create an auxiliary classifier, which captures richer visual variation, bridging text-image-pair classification to a more robust image-image-pair classification. Additionally, we exploit the diversity of diffusion-based synthetic samples to enhance prompt learning, providing greater information for semantic-visual alignment. Extensive experiments on five public benchmarks, including RESISC45 and Flowers102, and across three learning paradigms-UL, SSL, and TRZSL-demonstrate that AiR achieves substantial and consistent performance improvements over state-of-the-art unsupervised prompt learning methods.
Multimedia file fragment classification (MFFC) aims to identify file fragment types, e.g., image/video, audio, and text without system metadata. It is of vital importance in multimedia storage and communication. Existing MFFC methods typically treat fragments as 1D byte sequences and emphasize the relations between separate bytes (interbytes) for classification. However, the more informative relations inside bytes (intrabytes) are overlooked and seldom investigated. By looking inside bytes, the bit-level details of file fragments can be accessed, enabling a more accurate classification. Motivated by this, we first propose Byte2Image, a novel visual representation model that incorporates previously overlooked intrabyte information into file fragments and reinterprets these fragments as 2D grayscale images. This model involves a sliding byte window to reveal the intrabyte information and a rowwise stacking of intrabyte ngrams for embedding fragments into a 2D space. Thus, complex interbyte and intrabyte correlations can be mined simultaneously using powerful vision networks. Additionally, we propose an end-to-end dual-branch network ByteNet to enhance robust correlation mining and feature representation. ByteNet makes full use of the raw 1D byte sequence and the converted 2D image through a shallow byte branch feature extraction (BBFE) and a deep image branch feature extraction (IBFE) network. In particular, the BBFE, composed of a single fully-connected layer, adaptively recognizes the co-occurrence of several some specific bytes within the raw byte sequence, while the IBFE, built on a vision Transformer, effectively mines the complex interbyte and intrabyte correlations from the converted image. Experiments on the two representative benchmarks, including 14 cases, validate that our proposed method outperforms state-of-the-art approaches on different cases by up to 12.2%.
This work examines the vulnerability of multimodal (image + text) models to adversarial threats similar to those discussed in previous literature on unimodal (image- or text-only) models. We introduce realistic assumptions of partial model knowledge and access, and discuss how these assumptions differ from the standard "black-box"/"white-box" dichotomy common in current literature on adversarial attacks. Working under various levels of these "gray-box" assumptions, we develop new attack methodologies unique to multimodal classification and evaluate them on the Hateful Memes Challenge classification task. We find that attacking multiple modalities yields stronger attacks than unimodal attacks alone (inducing errors in up to 73% of cases), and that the unimodal image attacks on multimodal classifiers we explored were stronger than character-based text augmentation attacks (inducing errors on average in 45% and 30% of cases, respectively).
Image steganalysis is a technique for detecting data hidden in images. Recent research has shown the powerful capabilities of using convolutional neural networks (CNN) for image steganalysis. However, due to the particularity of steganographic signals, there are still few reliable CNN-based methods for applying steganalysis to images of arbitrary size. In this paper, we address this issue by exploring the possibility of exploiting a network for steganalyzing images of varying sizes without retraining its parameters. On the assumption that natural image noise is similar between different image sub-regions, we propose an end-to-end, deep learning, novel solution for distinguishing steganography images from normal images that provides satisfying performance. The proposed network first takes the image as the input, then identifies the relationships between the noise of different image sub-regions, and, finally, outputs the resulting classification based upon them. Our algorithm adopts a Siamese, CNN-based architecture, which consists of two symmetrical subnets with shared parameters, and contains three phases: preprocessing, feature extraction, and fusion/classification. To validate the network, we generated datasets composed of steganography images with multiple sizes and their corresponding normal images sourced from BOSSbase 1.01 and ALASKA #2. Experimental results produced by the data generated by various methods show that our proposed network is well-generalized and robust.
With the development of a series of Galaxy sky surveys in recent years, the observations increased rapidly, which makes the research of machine learning methods for galaxy image recognition a hot topic. Available automatic galaxy image recognition researches are plagued by the large differences in similarity between categories, the imbalance of data between different classes, and the discrepancy between the discrete representation of Galaxy classes and the essentially gradual changes from one morphological class to the adjacent class (DDRGC). These limitations have motivated several astronomers and machine learning experts to design projects with improved galaxy image recognition capabilities. Therefore, this paper proposes a novel learning method, ``Hierarchical Imbalanced data learning with Weighted sampling and Label smoothing" (HIWL). The HIWL consists of three key techniques respectively dealing with the above-mentioned three problems: (1) Designed a hierarchical galaxy classification model based on an efficient backbone network; (2) Utilized a weighted sampling scheme to deal with the imbalance problem; (3) Adopted a label smoothing technique to alleviate the DDRGC problem. We applied this method to galaxy photometric images from the Galaxy Zoo-The Galaxy Challenge, exploring the recognition of completely round smooth, in between smooth, cigar-shaped, edge-on and spiral. The overall classification accuracy is 96.32\%, and some superiorities of the HIWL are shown based on recall, precision, and F1-Score in comparing with some related works. In addition, we also explored the visualization of the galaxy image features and model attention to understand the foundations of the proposed scheme.
Multi-label radiography image classification has long been a topic of interest in neural networks research. In this paper, we intend to classify such images using convolution neural networks with novel localization techniques. We will use the chest x-ray images to detect thoracic diseases for this purpose. For accurate diagnosis, it is crucial to train the network with good quality images. But many chest X-ray images have irrelevant external objects like distractions created by faulty scans, electronic devices scanned next to lung region, scans inadvertently capturing bodily air etc. To address these, we propose a combination of localization and deep learning algorithms called LeDNet to predict thoracic diseases with higher accuracy. We identify and extract the lung region masks from chest x-ray images through localization. These masks are superimposed on the original X-ray images to create the mask overlay images. DenseNet-121 classification models are then used for feature selection to retrieve features of the entire chest X-ray images and the localized mask overlay images. These features are then used to predict disease classification. Our experiments involve comparing classification results obtained with original CheXpert images and mask overlay images. The comparison is demonstrated through accuracy and loss curve analyses.
Attributes act as intermediate representations that enable parameter sharing between classes, a must when training data is scarce. We propose to view attribute-based image classification as a label-embedding problem: each class is embedded in the space of attribute vectors. We introduce a function that measures the compatibility between an image and a label embedding. The parameters of this function are learned on a training set of labeled samples to ensure that, given an image, the correct classes rank higher than the incorrect ones. Results on the Animals With Attributes and Caltech-UCSD-Birds datasets show that the proposed framework outperforms the standard Direct Attribute Prediction baseline in a zero-shot learning scenario. Label embedding enjoys a built-in ability to leverage alternative sources of information instead of or in addition to attributes, such as e.g. class hierarchies or textual descriptions. Moreover, label embedding encompasses the whole range of learning settings from zero-shot learning to regular learning with a large number of labeled examples.
Vision Transformers (ViTs) have successfully been applied to image classification problems where large annotated datasets are available. On the other hand, when fewer annotations are available, such as in biomedical applications, image augmentation techniques like introducing image variations or combinations have been proposed. However, regarding ViT patch sampling, less has been explored outside grid-based strategies. In this work, we propose Random Vision Transformer Tokens (RaViTT), a random patch sampling strategy that can be incorporated into existing ViTs. We experimentally evaluated RaViTT for image classification, comparing it with a baseline ViT and state-of-the-art (SOTA) augmentation techniques in 4 datasets, including ImageNet-1k and CIFAR-100. Results show that RaViTT increases the accuracy of the baseline in all datasets and outperforms the SOTA augmentation techniques in 3 out of 4 datasets by a significant margin +1.23% to +4.32%. Interestingly, RaViTT accuracy improvements can be achieved even with fewer tokens, thus reducing the computational load of any ViT model for a given accuracy value.
We propose a novel framework for few-shot learning by leveraging large-scale vision-language models such as CLIP. Motivated by unimodal prototypical networks for few-shot learning, we introduce Proto-CLIP which utilizes image prototypes and text prototypes for few-shot learning. Specifically, Proto-CLIP adapts the image and text encoder embeddings from CLIP in a joint fashion using few-shot examples. The embeddings from the two encoders are used to compute the respective prototypes of image classes for classification. During adaptation, we propose aligning the image and text prototypes of the corresponding classes. Such alignment is beneficial for few-shot classification due to the reinforced contributions from both types of prototypes. Proto-CLIP has both training-free and fine-tuned variants. We demonstrate the effectiveness of our method by conducting experiments on benchmark datasets for few-shot learning, as well as in the real world for robot perception. The project page is available at https://irvlutd.github.io/Proto-CLIP
X-ray imaging is pivotal in medical diagnostics, offering non-invasive insights into a range of health conditions. Recently, vision-language models, such as the Contrastive Language-Image Pretraining (CLIP) model, have demonstrated potential in improving diagnostic accuracy by leveraging large-scale image-text datasets. However, since CLIP was not initially designed for medical images, several CLIP-like models trained specifically on medical images have been developed. Despite their enhanced performance, issues of fairness - particularly regarding demographic attributes - remain largely unaddressed. In this study, we perform a comprehensive fairness analysis of CLIP-like models applied to X-ray image classification. We assess their performance and fairness across diverse patient demographics and disease categories using zero-shot inference and various fine-tuning techniques, including Linear Probing, Multilayer Perceptron (MLP), Low-Rank Adaptation (LoRA), and full fine-tuning. Our results indicate that while fine-tuning improves model accuracy, fairness concerns persist, highlighting the need for further fairness interventions in these foundational models.
Computer-aided pathology diagnosis based on the classification of Whole Slide Image (WSI) plays an important role in clinical practice, and it is often formulated as a weakly-supervised Multiple Instance Learning (MIL) problem. Existing methods solve this problem from either a bag classification or an instance classification perspective. In this paper, we propose an end-to-end weakly supervised knowledge distillation framework (WENO) for WSI classification, which integrates a bag classifier and an instance classifier in a knowledge distillation framework to mutually improve the performance of both classifiers. Specifically, an attention-based bag classifier is used as the teacher network, which is trained with weak bag labels, and an instance classifier is used as the student network, which is trained using the normalized attention scores obtained from the teacher network as soft pseudo labels for the instances in positive bags. An instance feature extractor is shared between the teacher and the student to further enhance the knowledge exchange between them. In addition, we propose a hard positive instance mining strategy based on the output of the student network to force the teacher network to keep mining hard positive instances. WENO is a plug-and-play framework that can be easily applied to any existing attention-based bag classification methods. Extensive experiments on five datasets demonstrate the efficiency of WENO. Code is available at https://github.com/miccaiif/WENO.
Image generation abilities of text-to-image diffusion models have significantly advanced, yielding highly photo-realistic images from descriptive text and increasing the viability of leveraging synthetic images to train computer vision models. To serve as effective training data, generated images must be highly realistic while also sufficiently diverse within the support of the target data distribution. Yet, state-of-the-art conditional image generation models have been primarily optimized for creative applications, prioritizing image realism and prompt adherence over conditional diversity. In this paper, we investigate how to improve the diversity of generated images with the goal of increasing their effectiveness to train downstream image classification models, without fine-tuning the image generation model. We find that conditioning the generation process on an augmented real image and text prompt produces generations that serve as effective synthetic datasets for downstream training. Conditioning on real training images contextualizes the generation process to produce images that are in-domain with the real image distribution, while data augmentations introduce visual diversity that improves the performance of the downstream classifier. We validate augmentation-conditioning on a total of five established long-tail and few-shot image classification benchmarks and show that leveraging augmentations to condition the generation process results in consistent improvements over the state-of-the-art on the long-tailed benchmark and remarkable gains in extreme few-shot regimes of the remaining four benchmarks. These results constitute an important step towards effectively leveraging synthetic data for downstream training.
Cancer is a complex disease that provides various types of information depending on the scale of observation. While most tumor diagnostics are performed by observing histopathological slides, radiology images should yield additional knowledge towards the efficacy of cancer diagnostics. This work investigates a deep learning method combining whole slide images and magnetic resonance images to classify tumors. In particular, our solution comprises a powerful, generic and modular architecture for whole slide image classification. Experiments are prospectively conducted on the 2020 Computational Precision Medicine challenge, in a 3-classes unbalanced classification task. We report cross-validation (resp. validation) balanced-accuracy, kappa and f1 of 0.913, 0.897 and 0.951 (resp. 0.91, 0.90 and 0.94). For research purposes, including reproducibility and direct performance comparisons, our finale submitted models are usable off-the-shelf in a Docker image available at https://hub.docker.com/repository/docker/marvinler/cpm_2020_marvinler.
In this work, we introduce Mask-JEPA, a self-supervised learning framework tailored for mask classification architectures (MCA), to overcome the traditional constraints associated with training segmentation models. Mask-JEPA combines a Joint Embedding Predictive Architecture with MCA to adeptly capture intricate semantics and precise object boundaries. Our approach addresses two critical challenges in self-supervised learning: 1) extracting comprehensive representations for universal image segmentation from a pixel decoder, and 2) effectively training the transformer decoder. The use of the transformer decoder as a predictor within the JEPA framework allows proficient training in universal image segmentation tasks. Through rigorous evaluations on datasets such as ADE20K, Cityscapes and COCO, Mask-JEPA demonstrates not only competitive results but also exceptional adaptability and robustness across various training scenarios. The architecture-agnostic nature of Mask-JEPA further underscores its versatility, allowing seamless adaptation to various mask classification family.
Diabetic retinopathy (DR) is a leading cause of blindness worldwide, underscoring the importance of early detection for effective treatment. However, automated DR classification remains challenging due to variations in image quality, class imbalance, and pixel-level similarities that hinder model training. To address these issues, we propose a robust preprocessing pipeline incorporating image cropping, Contrast-Limited Adaptive Histogram Equalization (CLAHE), and targeted data augmentation to improve model generalization and resilience. Our approach leverages the Swin Transformer, which utilizes hierarchical token processing and shifted window attention to efficiently capture fine-grained features while maintaining linear computational complexity. We validate our method on the Aptos and IDRiD datasets for multi-class DR classification, achieving accuracy rates of 89.65% and 97.40%, respectively. These results demonstrate the effectiveness of our model, particularly in detecting early-stage DR, highlighting its potential for improving automated retinal screening in clinical settings.
In this letter, we propose a multitask deep learning method for classification of multiple hyperspectral data in a single training. Deep learning models have achieved promising results on hyperspectral image classification, but their performance highly rely on sufficient labeled samples, which are scarce on hyperspectral images. However, samples from multiple data sets might be sufficient to train one deep learning model, thereby improving its performance. To do so, we trained an identical feature extractor for all data, and the extracted features were fed into corresponding Softmax classifiers. Spectral knowledge was introduced to ensure that the shared features were similar across domains. Four hyperspectral data sets were used in the experiments. We achieved higher classification accuracies on three data sets (Pavia University, Pavia Center, and Indian Pines) and competitive results on the Salinas Valley data compared with the baseline. Spectral knowledge was useful to prevent the deep network from overfitting when the data shared similar spectral response. The proposed method tested on two deep CNNs successfully shows its ability to utilize samples from multiple data sets and enhance networks' performance.
The state of an object is an important piece of knowledge in robotics applications. States and objects are intertwined together, meaning that object information can help recognize the state of an image and vice versa. This paper addresses the state identification problem in cooking related images and uses state and object predictions together to improve the classification accuracy of objects and their states from a single image. The pipeline presented in this paper includes a CNN with a double classification layer and the Concept-Net language knowledge graph on top. The language knowledge creates a semantic likelihood between objects and states. The resulting object and state confidences from the deep architecture are used together with object and state relatedness estimates from a language knowledge graph to produce marginal probabilities for objects and states. The marginal probabilities and confidences of objects (or states) are fused together to improve the final object (or state) classification results. Experiments on a dataset of cooking objects show that using a language knowledge graph on top of a deep neural network effectively enhances object and state classification.
Cardiac Magnetic Resonance Imaging is commonly used for the assessment of the cardiac anatomy and function. The delineations of left and right ventricle blood pools and left ventricular myocardium are important for the diagnosis of cardiac diseases. Unfortunately, the movement of a patient during the CMR acquisition procedure may result in motion artifacts appearing in the final image. Such artifacts decrease the diagnostic quality of CMR images and force redoing of the procedure. In this paper, we present a Multi-task Swin UNEt TRansformer network for simultaneous solving of two tasks in the CMRxMotion challenge: CMR segmentation and motion artifacts classification. We utilize both segmentation and classification as a multi-task learning approach which allows us to determine the diagnostic quality of CMR and generate masks at the same time. CMR images are classified into three diagnostic quality classes, whereas, all samples with non-severe motion artifacts are being segmented. Ensemble of five networks trained using 5-Fold Cross-validation achieves segmentation performance of DICE coefficient of 0.871 and classification accuracy of 0.595.
Data augmentation is a central component of joint embedding self-supervised learning (SSL). Approaches that work for natural images may not always be effective in medical imaging tasks. This study systematically investigated the impact of data augmentation and preprocessing strategies in SSL for lung ultrasound. Three data augmentation pipelines were assessed: (1) a baseline pipeline commonly used across imaging domains, (2) a novel semantic-preserving pipeline designed for ultrasound, and (3) a distilled set of the most effective transformations from both pipelines. Pretrained models were evaluated on multiple classification tasks: B-line detection, pleural effusion detection, and COVID-19 classification. Experiments revealed that semantics-preserving data augmentation resulted in the greatest performance for COVID-19 classification - a diagnostic task requiring global image context. Cropping-based methods yielded the greatest performance on the B-line and pleural effusion object classification tasks, which require strong local pattern recognition. Lastly, semantics-preserving ultrasound image preprocessing resulted in increased downstream performance for multiple tasks. Guidance regarding data augmentation and preprocessing strategies was synthesized for practitioners working with SSL in ultrasound.
Deep Learning approaches in dermatological image classification have shown promising results, yet the field faces significant methodological challenges that impede proper evaluation. This paper presents a dual contribution: first, a systematic analysis of current methodological practices in skin disease classification research, revealing substantial inconsistencies in data preparation, augmentation strategies, and performance reporting; second, a comprehensive training and evaluation framework demonstrated through experiments with the DINOv2-Large vision transformer across three benchmark datasets (HAM10000, DermNet, ISIC Atlas). The analysis identifies concerning patterns, including pre-split data augmentation and validation-based reporting, potentially leading to overestimated metrics, while highlighting the lack of unified methodology standards. The experimental results demonstrate DINOv2's performance in skin disease classification, achieving macro-averaged F1-scores of 0.85 (HAM10000), 0.71 (DermNet), and 0.84 (ISIC Atlas). Attention map analysis reveals critical patterns in the model's decision-making, showing sophisticated feature recognition in typical presentations but significant vulnerabilities with atypical cases and composite images. Our findings highlight the need for standardized evaluation protocols and careful implementation strategies in clinical settings. We propose comprehensive methodological recommendations for model development, evaluation, and clinical deployment, emphasizing rigorous data preparation, systematic error analysis, and specialized protocols for different image types. To promote reproducibility, we provide our implementation code through GitHub. This work establishes a foundation for rigorous evaluation standards in dermatological image classification and provides insights for responsible AI implementation in clinical dermatology.