Radar-based wellness monitoring is becoming an effective measurement to provide accurate vital signs in a contactless manner, but data scarcity retards the related research on deep-learning-based methods. Data augmentation is commonly used to enrich the dataset by modifying the existing data, but most augmentation techniques can only couple with classification tasks. To enable the augmentation for regression tasks, this research proposes a spectrogram augmentation method, Horcrux, for radar-based cardiac feature monitoring (e.g., heartbeat detection, electrocardiogram reconstruction) with both classification and regression tasks involved. The proposed method is designed to increase the diversity of input samples while the augmented spectrogram is still faithful to the original ground truth vital sign. In addition, Horcrux proposes to inject zero values in specific areas to enhance the awareness of the deep learning model on subtle cardiac features, improving the performance for the limited dataset. Experimental result shows that Horcrux achieves an overall improvement of 16.20% in cardiac monitoring and has the potential to be extended to other spectrogram-based tasks. The code will be released upon publication.
Distributed acoustic sensing (DAS) technology represents an innovative fiber-optic-based sensing methodology that enables real-time acoustic signal monitoring through the detection of minute perturbations along optical fibers. This sensing approach offers compelling advantages, including extensive measurement ranges, exceptional spatial resolution, and an expansive dynamic measurement spectrum. The integration of machine learning (ML) paradigms presents transformative potential for DAS technology, encompassing critical domains such as data augmentation, sophisticated preprocessing techniques, and advanced acoustic event classification and recognition. By leveraging ML algorithms, DAS systems can transition from traditional data processing methodologies to more automated and intelligent analytical frameworks. The computational intelligence afforded by ML-enhanced DAS technologies facilitates unprecedented monitoring capabilities across diverse critical infrastructure sectors. Particularly noteworthy are the technology's applications in transportation infrastructure, energy management systems, and Natural disaster monitoring frameworks, where the precision of data acquisition and the reliability of intelligent decision-making mechanisms are paramount. This research critically examines the comparative performance characteristics of classical machine learning methodologies and state-of-the-art deep learning models in the context of DAS data recognition and interpretation, offering comprehensive insights into the evolving landscape of intelligent sensing technologies.
During criminal investigations, images of persons of interest directly influence the success of identification procedures. However, law enforcement agencies often face challenges related to the scarcity of high-quality images or their obsolescence, which can affect the accuracy and success of people searching processes. This paper introduces a novel forensic mugshot augmentation framework aimed at addressing these limitations. Our approach enhances the identification probability of individuals by generating additional, high-quality images through customizable data augmentation techniques, while maintaining the biometric integrity and consistency of the original data. Several experimental results show that our method significantly improves identification accuracy and robustness across various forensic scenarios, demonstrating its effectiveness as a trustworthy tool law enforcement applications. Index Terms: Digital Forensics, Person re-identification, Feature extraction, Data augmentation, Visual-Language models.
In the computer vision community, the preference for pre-training visual models has largely shifted toward sRGB images due to their ease of acquisition and compact storage. However, camera RAW images preserve abundant physical details across diverse real-world scenarios. Despite this, most existing visual perception methods that utilize RAW data directly integrate image signal processing (ISP) stages with subsequent network modules, often overlooking potential synergies at the model level. Building on recent advances in adapter-based methodologies in both NLP and computer vision, we propose RAW-Adapter, a novel framework that incorporates learnable ISP modules as input-level adapters to adjust RAW inputs. At the same time, it employs model-level adapters to seamlessly bridge ISP processing with high-level downstream architectures. Moreover, RAW-Adapter serves as a general framework applicable to various computer vision frameworks. Furthermore, we introduce RAW-Bench, which incorporates 17 types of RAW-based common corruptions, including lightness degradations, weather effects, blurriness, camera imaging degradations, and variations in camera color response. Using this benchmark, we systematically compare the performance of RAW-Adapter with state-of-the-art (SOTA) ISP methods and other RAW-based high-level vision algorithms. Additionally, we propose a RAW-based data augmentation strategy to further enhance RAW-Adapter's performance and improve its out-of-domain (OOD) generalization ability. Extensive experiments substantiate the effectiveness and efficiency of RAW-Adapter, highlighting its robust performance across diverse scenarios.
Semi-supervised object detection (SSOD) based on pseudo-labeling significantly reduces dependence on large labeled datasets by effectively leveraging both labeled and unlabeled data. However, real-world applications of SSOD often face critical challenges, including class imbalance, label noise, and labeling errors. We present an in-depth analysis of SSOD under real-world conditions, uncovering causes of suboptimal pseudo-labeling and key trade-offs between label quality and quantity. Based on our findings, we propose four building blocks that can be seamlessly integrated into an SSOD framework. Rare Class Collage (RCC): a data augmentation method that enhances the representation of rare classes by creating collages of rare objects. Rare Class Focus (RCF): a stratified batch sampling strategy that ensures a more balanced representation of all classes during training. Ground Truth Label Correction (GLC): a label refinement method that identifies and corrects false, missing, and noisy ground truth labels by leveraging the consistency of teacher model predictions. Pseudo-Label Selection (PLS): a selection method for removing low-quality pseudo-labeled images, guided by a novel metric estimating the missing detection rate while accounting for class rarity. We validate our methods through comprehensive experiments on autonomous driving datasets, resulting in up to 6% increase in SSOD performance. Overall, our investigation and novel, data-centric, and broadly applicable building blocks enable robust and effective SSOD in complex, real-world scenarios. Code is available at https://mos-ks.github.io/publications.
The aim of this work is to explore the potential of pre-trained vision-language models, e.g. Vision Transformers (ViT), enhanced with advanced data augmentation strategies for the detection of AI-generated images. Our approach leverages a fine-tuned ViT model trained on the Defactify-4.0 dataset, which includes images generated by state-of-the-art models such as Stable Diffusion 2.1, Stable Diffusion XL, Stable Diffusion 3, DALL-E 3, and MidJourney. We employ perturbation techniques like flipping, rotation, Gaussian noise injection, and JPEG compression during training to improve model robustness and generalisation. The experimental results demonstrate that our ViT-based pipeline achieves state-of-the-art performance, significantly outperforming competing methods on both validation and test datasets.
We present a domain adaptation of video diffusion models to generate highly realistic time-lapse microscopy videos of cell division in HeLa cells. Although state-of-the-art generative video models have advanced significantly for natural videos, they remain underexplored in microscopy domains. To address this gap, we fine-tune a pretrained video diffusion model on microscopy-specific sequences, exploring three conditioning strategies: (1) text prompts derived from numeric phenotypic measurements (e.g., proliferation rates, migration speeds, cell-death frequencies), (2) direct numeric embeddings of phenotype scores, and (3) image-conditioned generation, where an initial microscopy frame is extended into a complete video sequence. Evaluation using biologically meaningful morphological, proliferation, and migration metrics demonstrates that fine-tuning substantially improves realism and accurately captures critical cellular behaviors such as mitosis and migration. Notably, the fine-tuned model also generalizes beyond the training horizon, generating coherent cell dynamics even in extended sequences. However, precisely controlling specific phenotypic characteristics remains challenging, highlighting opportunities for future work to enhance conditioning methods. Our results demonstrate the potential for domain-specific fine-tuning of generative video models to produce biologically plausible synthetic microscopy data, supporting applications such as in-silico hypothesis testing and data augmentation.
Deep learning has made significant progress in addressing challenges in various fields including computational pathology (CPath). However, due to the complexity of the domain shift problem, the performance of existing models will degrade, especially when it comes to multi-domain or cross-domain tasks. In this paper, we propose a Test-time style transfer (T3s) that uses a bidirectional mapping mechanism to project the features of the source and target domains into a unified feature space, enhancing the generalization ability of the model. To further increase the style expression space, we introduce a Cross-domain style diversification module (CSDM) to ensure the orthogonality between style bases. In addition, data augmentation and low-rank adaptation techniques are used to improve feature alignment and sensitivity, enabling the model to adapt to multi-domain inputs effectively. Our method has demonstrated effectiveness on three unseen datasets.
Human activity recognition (HAR) with deep learning models relies on large amounts of labeled data, often challenging to obtain due to associated cost, time, and labor. Self-supervised learning (SSL) has emerged as an effective approach to leverage unlabeled data through pretext tasks, such as masked reconstruction and multitask learning with signal processing-based data augmentations, to pre-train encoder models. However, such methods are often derived from computer vision approaches that disregard physical mechanisms and constraints that govern wearable sensor data and the phenomena they reflect. In this paper, we propose a physics-informed multi-task pre-training (PIM) framework for IMU-based HAR. PIM generates pre-text tasks based on the understanding of basic physical aspects of human motion: including movement speed, angles of movement, and symmetry between sensor placements. Given a sensor signal, we calculate corresponding features using physics-based equations and use them as pretext tasks for SSL. This enables the model to capture fundamental physical characteristics of human activities, which is especially relevant for multi-sensor systems. Experimental evaluations on four HAR benchmark datasets demonstrate that the proposed method outperforms existing state-of-the-art methods, including data augmentation and masked reconstruction, in terms of accuracy and F1 score. We have observed gains of almost 10\% in macro f1 score and accuracy with only 2 to 8 labeled examples per class and up to 3% when there is no reduction in the amount of training data.
Contrastive learning -- a modern approach to extract useful representations from unlabeled data by training models to distinguish similar samples from dissimilar ones -- has driven significant progress in foundation models. In this work, we develop a new theoretical framework for analyzing data augmentation-based contrastive learning, with a focus on SimCLR as a representative example. Our approach is based on the concept of \emph{approximate sufficient statistics}, which we extend beyond its original definition in \cite{oko2025statistical} for contrastive language-image pretraining (CLIP) using KL-divergence. We generalize it to equivalent forms and general f-divergences, and show that minimizing SimCLR and other contrastive losses yields encoders that are approximately sufficient. Furthermore, we demonstrate that these near-sufficient encoders can be effectively adapted to downstream regression and classification tasks, with performance depending on their sufficiency and the error induced by data augmentation in contrastive learning. Concrete examples in linear regression and topic classification are provided to illustrate the broad applicability of our results.
Deep learning has emerged as the predominant solution for classifying medical images. We intend to apply these developments to the ultra-widefield (UWF) retinal imaging dataset. Since UWF images can accurately diagnose various retina diseases, it is very important to clas sify them accurately and prevent them with early treatment. However, processing images manually is time-consuming and labor-intensive, and there are two challenges to automating this process. First, high perfor mance usually requires high computational resources. Artificial intelli gence medical technology is better suited for places with limited medical resources, but using high-performance processing units in such environ ments is challenging. Second, the problem of the accuracy of colour fun dus photography (CFP) methods. In general, the UWF method provides more information for retinal diagnosis than the CFP method, but most of the research has been conducted based on the CFP method. Thus, we demonstrate that these problems can be efficiently addressed in low performance units using methods such as strategic data augmentation and model ensembles, which balance performance and computational re sources while utilizing UWF images.
Left ventricular ejection fraction (LVEF) is a critical metric for assessing cardiac function, widely used in diagnosing heart failure and guiding clinical decisions. Despite its importance, conventional LVEF estimation remains time-consuming and operator-dependent. Recent deep learning advancements have enhanced automation, yet many existing models are computationally demanding, hindering their feasibility for real-time clinical applications. Additionally, the interplay between spatial and temporal features is crucial for accurate estimation but is often overlooked. In this work, we propose Echo-E$^3$Net, an efficient Endo-Epi spatio-temporal network tailored for LVEF estimation. Our method introduces the Endo-Epi Cardial Border Detector (E$^2$CBD) module, which enhances feature extraction by leveraging spatial and temporal landmark cues. Complementing this, the Endo-Epi Feature Aggregator (E$^2$FA) distills statistical descriptors from backbone feature maps, refining the final EF prediction. These modules, along with a multi-component loss function tailored to align with the clinical definition of EF, collectively enhance spatial-temporal representation learning, ensuring robust and efficient EF estimation. We evaluate Echo-E$^3$Net on the EchoNet-Dynamic dataset, achieving a RMSE of 5.15 and an R$^2$ score of 0.82, setting a new benchmark in efficiency with 6.8 million parameters and only 8.49G Flops. Our model operates without pre-training, data augmentation, or ensemble methods, making it well-suited for real-time point-of-care ultrasound (PoCUS) applications. Our Code is publicly available on~\href{https://github.com/moeinheidari7829/Echo-E3Net}{\textcolor{magenta}{GitHub}}.
Conditional human animation traditionally animates static reference images using pose-based motion cues extracted from video data. However, these video-derived cues often suffer from low temporal resolution, motion blur, and unreliable performance under challenging lighting conditions. In contrast, event cameras inherently provide robust and high temporal-resolution motion information, offering resilience to motion blur, low-light environments, and exposure variations. In this paper, we propose EvAnimate, the first method leveraging event streams as robust and precise motion cues for conditional human image animation. Our approach is fully compatible with diffusion-based generative models, enabled by encoding asynchronous event data into a specialized three-channel representation with adaptive slicing rates and densities. High-quality and temporally coherent animations are achieved through a dual-branch architecture explicitly designed to exploit event-driven dynamics, significantly enhancing performance under challenging real-world conditions. Enhanced cross-subject generalization is further achieved using specialized augmentation strategies. To facilitate future research, we establish a new benchmarking, including simulated event data for training and validation, and a real-world event dataset capturing human actions under normal and challenging scenarios. The experiment results demonstrate that EvAnimate achieves high temporal fidelity and robust performance in scenarios where traditional video-derived cues fall short.
Accurately predicting click-through rates (CTR) under stringent privacy constraints poses profound challenges, particularly when user-item interactions are sparse and fragmented across domains. Conventional cross-domain CTR (CCTR) methods frequently assume homogeneous feature spaces and rely on centralized data sharing, neglecting complex inter-domain discrepancies and the subtle trade-offs imposed by privacy-preserving protocols. Here, we present Federated Cross-Domain CTR Prediction with Large Language Model Augmentation (FedCCTR-LM), a federated framework engineered to address these limitations by synchronizing data augmentation, representation disentanglement, and adaptive privacy protection. Our approach integrates three core innovations. First, the Privacy-Preserving Augmentation Network (PrivAugNet) employs large language models to enrich user and item representations and expand interaction sequences, mitigating data sparsity and feature incompleteness. Second, the Independent Domain-Specific Transformer with Contrastive Learning (IDST-CL) module disentangles domain-specific and shared user preferences, employing intra-domain representation alignment (IDRA) and crossdomain representation disentanglement (CDRD) to refine the learned embeddings and enhance knowledge transfer across domains. Finally, the Adaptive Local Differential Privacy (AdaLDP) mechanism dynamically calibrates noise injection to achieve an optimal balance between rigorous privacy guarantees and predictive accuracy. Empirical evaluations on four real-world datasets demonstrate that FedCCTR-LM substantially outperforms existing baselines, offering robust, privacy-preserving, and generalizable cross-domain CTR prediction in heterogeneous, federated environments.
Deepfakes have become a universal and rapidly intensifying concern of generative AI across various media types such as images, audio, and videos. Among these, audio deepfakes have been of particular concern due to the ease of high-quality voice synthesis and distribution via platforms such as social media and robocalls. Consequently, detecting audio deepfakes plays a critical role in combating the growing misuse of AI-synthesized speech. However, real-world scenarios often introduce various audio corruptions, such as noise, modification, and compression, that may significantly impact detection performance. This work systematically evaluates the robustness of 10 audio deepfake detection models against 16 common corruptions, categorized into noise perturbation, audio modification, and compression. Using both traditional deep learning models and state-of-the-art foundation models, we make four unique observations. First, our findings show that while most models demonstrate strong robustness to noise, they are notably more vulnerable to modifications and compression, especially when neural codecs are applied. Second, speech foundation models generally outperform traditional models across most scenarios, likely due to their self-supervised learning paradigm and large-scale pre-training. Third, our results show that increasing model size improves robustness, albeit with diminishing returns. Fourth, we demonstrate how targeted data augmentation during training can enhance model resilience to unseen perturbations. A case study on political speech deepfakes highlights the effectiveness of foundation models in achieving high accuracy under real-world conditions. These findings emphasize the importance of developing more robust detection frameworks to ensure reliability in practical deployment settings.
Patients are increasingly turning to online health Q&A communities for social support to improve their well-being. However, when this support received does not align with their specific needs, it may prove ineffective or even detrimental. This necessitates a model capable of identifying the social support needs in questions. However, training such a model is challenging due to the scarcity and class imbalance issues of labeled data. To overcome these challenges, we follow the computational design science paradigm to develop a novel framework, Hybrid Approach for SOcial Support need classification (HA-SOS). HA-SOS integrates an answer-enhanced semi-supervised learning approach, a text data augmentation technique leveraging large language models (LLMs) with reliability- and diversity-aware sample selection mechanism, and a unified training process to automatically label social support needs in questions. Extensive empirical evaluations demonstrate that HA-SOS significantly outperforms existing question classification models and alternative semi-supervised learning approaches. This research contributes to the literature on social support, question classification, semi-supervised learning, and text data augmentation. In practice, our HA-SOS framework facilitates online Q&A platform managers and answerers to better understand users' social support needs, enabling them to provide timely, personalized answers and interventions.
Resolution generalization in image generation tasks enables the production of higher-resolution images with lower training resolution overhead. However, a significant challenge in resolution generalization, particularly in the widely used Diffusion Transformers, lies in the mismatch between the positional encodings encountered during testing and those used during training. While existing methods have employed techniques such as interpolation, extrapolation, or their combinations, none have fully resolved this issue. In this paper, we propose a novel two-dimensional randomized positional encodings (RPE-2D) framework that focuses on learning positional order of image patches instead of the specific distances between them, enabling seamless high- and low-resolution image generation without requiring high- and low-resolution image training. Specifically, RPE-2D independently selects positions over a broader range along both the horizontal and vertical axes, ensuring that all position encodings are trained during the inference phase, thus improving resolution generalization. Additionally, we propose a random data augmentation technique to enhance the modeling of position order. To address the issue of image cropping caused by the augmentation, we introduce corresponding micro-conditioning to enable the model to perceive the specific cropping patterns. On the ImageNet dataset, our proposed RPE-2D achieves state-of-the-art resolution generalization performance, outperforming existing competitive methods when trained at a resolution of $256 \times 256$ and inferred at $384 \times 384$ and $512 \times 512$, as well as when scaling from $512 \times 512$ to $768 \times 768$ and $1024 \times 1024$. And it also exhibits outstanding capabilities in low-resolution image generation, multi-stage training acceleration and multi-resolution inheritance.
Semi-supervised crowd counting is crucial for addressing the high annotation costs of densely populated scenes. Although several methods based on pseudo-labeling have been proposed, it remains challenging to effectively and accurately utilize unlabeled data. In this paper, we propose a novel framework called Taste More Taste Better (TMTB), which emphasizes both data and model aspects. Firstly, we explore a data augmentation technique well-suited for the crowd counting task. By inpainting the background regions, this technique can effectively enhance data diversity while preserving the fidelity of the entire scenes. Secondly, we introduce the Visual State Space Model as backbone to capture the global context information from crowd scenes, which is crucial for extremely crowded, low-light, and adverse weather scenarios. In addition to the traditional regression head for exact prediction, we employ an Anti-Noise classification head to provide less exact but more accurate supervision, since the regression head is sensitive to noise in manual annotations. We conduct extensive experiments on four benchmark datasets and show that our method outperforms state-of-the-art methods by a large margin. Code is publicly available on https://github.com/syhien/taste_more_taste_better.
Despite the recent success of image-text contrastive models like CLIP and SigLIP, these models often struggle with vision-centric tasks that demand high-fidelity image understanding, such as counting, depth estimation, and fine-grained object recognition. These models, by performing language alignment, tend to prioritize high-level semantics over visual understanding, weakening their image understanding. On the other hand, vision-focused models are great at processing visual information but struggle to understand language, limiting their flexibility for language-driven tasks. In this work, we introduce TULIP, an open-source, drop-in replacement for existing CLIP-like models. Our method leverages generative data augmentation, enhanced image-image and text-text contrastive learning, and image/text reconstruction regularization to learn fine-grained visual features while preserving global semantic alignment. Our approach, scaling to over 1B parameters, outperforms existing state-of-the-art (SOTA) models across multiple benchmarks, establishing a new SOTA zero-shot performance on ImageNet-1K, delivering up to a $2\times$ enhancement over SigLIP on RxRx1 in linear probing for few-shot classification, and improving vision-language models, achieving over $3\times$ higher scores than SigLIP on MMVP. Our code/checkpoints are available at https://tulip-berkeley.github.io
Efforts to address declining accuracy as a result of data shifts often involve various data-augmentation strategies. Adversarial training is one such method, designed to improve robustness to worst-case distribution shifts caused by adversarial examples. While this method can improve robustness, it may also hinder generalization to clean examples and exacerbate performance imbalances across different classes. This paper explores the impact of adversarial training on both overall and class-specific performance, as well as its spill-over effects. We observe that enhanced labeling during training boosts adversarial robustness by 53.50% and mitigates class imbalances by 5.73%, leading to improved accuracy in both clean and adversarial settings compared to standard adversarial training.