Automation in construction is essential for reducing costs and human errors in large-scale projects. We approach the construction progress monitoring from the aspect of detecting changes in construction sites. As construction buildings continue to evolve in geometry and appearance over time, change detection need to be performed from arbitrary camera viewpoints. This necessitates developing 2D Change Detection (2DCD) algorithms that operate robustly across diverse camera perspectives at construction sites. While developing and evaluating such systems is data-intensive, no open-source benchmark dataset exists at the intersection of 2D change detection and construction automation research. Data collection using Unmanned Aerial Vehicles (UAVs) is gaining its popularity in outdoor large-scale surveying. However, in active construction sites conducting drone missions equipped with high-end sensors imposes safety concerns. Flight trajectory and collected camera viewpoints can be significantly limited. To address this critical gap, we introduce iVISION-2DCD, a large-scale synthetically generated dataset from dense LiDAR point clouds with photorealistic input images and accurate ground truth annotations. Our dataset formally defines the problem of viewpoint-robust 2DCD at construction sites and captures the inherent complexities of real-world deployment. In this paper, we present our systematic methodology for synthetic data generation, developing novel view synthesis techniques to overcome bi-temporal alignment and viewpoint diversity challenges, and implementing semi-automated semantic segmentation with change label generation while preserving challenging real-world cases. Benchmark evaluations using state-of-the-art 2DCD algorithms demonstrate that iVISION-2DCD poses novel research challenges for the computer vision and robotics communities.

Vision transformers have shown remarkable performance in vision tasks, but enabling them for accessible and real-time use is still challenging. Quantization reduces memory and inference costs at the risk of performance loss. Strides have been made to mitigate low precision issues mainly by understanding in-distribution (ID) task behaviour, but the attention mechanism may provide insight on quantization attributes by exploring out-of-distribution (OOD) situations. We investigate the behaviour of quantized small-variant popular vision transformers (DeiT, DeiT3, and ViT) on common OOD datasets. ID analyses show the initial instabilities of 4-bit models, particularly of those trained on the larger ImageNet-22k, as the strongest FP32 model, DeiT3, sharply drop 17

Tuberculosis (TB) remains a persistent global health challenge, particularly in resource-constrained and remote regions where healthcare access is limited. Despite being both curable and preventable, TB continues to cause significant morbidity and mortality, emphasizing the urgent need for early detection and large-scale screening of at-risk populations. In recent years, artificial intelligence (AI), and more specifically deep learning, has emerged as a transformative tool for automating medical image analysis and supporting clinical decision-making. However, the reliability, robustness, and security of these AI solutions are critical concerns, as their vulnerability to adversarial attacks poses serious risks in safety-critical healthcare environments. This study systematically investigates the adversarial robustness of deep learning models for TB screening using chest X-ray images. A diverse set of convolutional and transformer-based architectures is evaluated under a range of white-box and black-box adversarial attack scenarios. Experimental results reveal that both model families exhibit significant performance degradation when subjected to adversarial perturbations. Our findings also suggest that leveraging a feature-encoder–based defense framework can significantly improve each model’s capability to handle adversarial attacks. This approach allows the model to maintain high diagnostic accuracy on unperturbed images while abstaining from unreliable predictions on potentially adversarial samples.

Tuberculosis remains a critical global health issue, particularly in resource-limited and remote areas. Early detection is vital for treatment, yet the lack of skilled radiologists underscores the need for artificial intelligence (AI)-driven screening tools. Developing reliable AI models is challenging due to the necessity for large, high-quality datasets, which are costly to obtain. To tackle this, we propose a teacher–student framework which enhances both disease and symptom detection on chest X-rays by integrating two supervised heads and a self-supervised head. Our model achieves an accuracy of 98.85

Unsupervised domain adaptation for semantic segmentation (DASS) aims to transfer knowledge from a label-rich source domain to a target domain with no labels. Two key approaches in DASS are (1) vision-only approaches using masking or multi-resolution crops, and (2) language-based approaches that use generic class-wise prompts informed by target domain (e.g. "a snowy photo of a class"). However, the former is susceptible to noisy pseudo-labels that are biased to the source domain. The latter does not fully capture the intricate spatial relationships of objects – key for dense prediction tasks. To this end, we propose LangDA. LangDA addresses these challenges by, first, learning contextual relationships between objects via VLM-generated scene descriptions (e.g. "a pedestrian is on the sidewalk, and the street is lined with buildings."). Second, LangDA aligns the entire image features with text representation of this context-aware scene caption and learns generalized representations via text. With this, LangDA sets the new state-of-the-art across three DASS benchmarks, outperforming existing methods by 2.6 3.9

Video Scene Graph Generation (VidSGG) is an important topic in understanding dynamic kitchen environments. Current models for VidSGG require extensive training to produce scene graphs. Recently, Vision Language Models (VLM) and Vision Foundation Models (VFM) have demonstrated impressive zero-shot capabilities in a variety of tasks. However, VLMs like Gemini struggle with the dynamics for VidSGG, failing to maintain stable object identities across frames. To overcome this limitation, we propose SAMJAM, a zero-shot pipeline that combines SAM2's temporal tracking with Gemini's semantic understanding. SAM2 also improves upon Gemini's object grounding by producing more accurate bounding boxes. In our method, we first prompt Gemini to generate a frame-level scene graph. Then, we employ a matching algorithm to map each object in the scene graph with a SAM2-generated or SAM2-propagated mask, producing a temporally-consistent scene graph in dynamic environments. Finally, we repeat this process again in each of the following frames. We empirically demonstrate that SAMJAM outperforms Gemini by 8.33% in mean recall on the EPIC-KITCHENS and EPIC-KITCHENS-100 datasets.

Segmenting medical images to identify lesions, organs and other areas of interest is crucial for diagnosis and treatment decisions. Traditionally, segmentation is accomplished through manual tools or using automated task-specific neural network models. A promising alternative solution to this problem is to create general-purpose models for segment anything in medical images, such as MedSAM [8]. These foundation models can segment regions across a multitude of modalities, at levels comparable to task-specific models. However, these models are often large and computationally expensive, preventing them from being used in clinical settings where they lack dedicated GPUs. We propose an efficient model for the segment anything in medical images problem, RepViT-MedSAM, created from a two step training process. First, the image encoder of MedSAM is distilled into a more efficient RepViT feature detector using aggressively augmented medical images. Then the entire end-to-end model, with the prompt encoder and mask decoder, is fine-tuned using ground truth masks and MedSAM's predictions. On the test set, RepViT-MedSAM surpasses the performance of baseline MedSAM in performance and efficiency, achieving an average Dice Similarity Coefficient (DSC) of 0.8528, an average Normalized Surface Distance (NSD) of 0.8666, taking a total execution time of 195 s, and ranking 12/23 among other contestants. RepViT-SAM offers a promising solution for real-world medical image segmentation with its efficiency and accuracy. The code for this project is available at https://github.com/ icecap360/TurboMedSAM.

Structured scene representations are a core component of embodied agents, helping to consolidate raw sensory streams into readable, modular, and searchable formats. Due to their high computational overhead, many approaches build such representations in advance of the task. However, when the task specifications change, such static approaches become inadequate as they may miss key objects, spatial relations, and details. We introduce GraphPad, a modifiable structured memory that an agent can tailor to the needs of the task through API calls. It comprises a mutable scene graph representing the environment, a navigation log indexing frame-by-frame content, and a scratchpad for task-specific notes. Together, GraphPad serves as a dynamic workspace that remains complete, current, and aligned with the agent's immediate understanding of the scene and its task. On the OpenEQA benchmark, GraphPad attains 55.3 +3.0 model, while operating with five times fewer input frames. These results show that allowing online, language-driven refinement of 3-D memory yields more informative representations without extra training or data collection.

Objective: This study introduces a physics-informed diffusion model (PIDM) for super-resolution (SR) reconstruction of optical coherence tomography (OCT) data. Methods: An optimization framework was developed for maximizing the likelihood of observing an OCT image in the dataset, given the super-resolved reconstruction from a physics-informed diffusion model (PIDM) that reverses the degradations in OCT images. The image degradations were modeled as a serialization of three processes accounting for the effects of defocus, speckle noise, and digital sampling in OCT images. An analytical model for light-propagation model and a statistical model for speckle noise were derived based on the physical properties of the OCT setup. These models were then integrated with a diffusion model to reverse the degradations caused by defocus blur and digital sampling, minimizing susceptibility to noise and defocus-induced artifacts. Results: The proposed method was employed for reconstructing images of a standard resolution target, plant tissue, and in vivo human cornea, using the complex OCT data acquired with a line-scan OCT (LS-OCT) system. The results from the PIDM exhibit improved sharpness and contrast compared to the images resulting from a few baseline methods such as standalone super-resolution using DM. Conclusion: Complementing DM with the physics of OCT could be a viable solution for obtaining high-fidelity SR reconstruction of OCT images. Significance: This work harnesses the power of diffusion models for super-resolution in OCT images. Such development could potentially enhance cellular-resolution OCT imaging of ophthalmic tissues, where high-fidelity images are crucial for accurate diagnosis.

Prostate cancer (PCa) is the most prevalent cancer among men in the United States, accounting for nearly 300,000 cases, 29 total deaths in 2024. Traditional screening methods such as prostate-specific antigen (PSA) testing and magnetic resonance imaging (MRI) have been pivotal in diagnosis, but have faced limitations in specificity and generalizability. In this paper, we explore the potential of enhancing PCa lesion segmentation using a novel MRI modality called synthetic correlated diffusion imaging (CDI^s). We employ several state-of-the-art deep learning models, including U-Net, SegResNet, Swin UNETR, Attention U-Net, and LightM-UNet, to segment PCa lesions from a 200 CDI^s patient cohort. We find that SegResNet achieved superior segmentation performance with a Dice-Sorensen coefficient (DSC) of 76.68 ± 0.8. Notably, the Attention U-Net, while slightly less accurate (DSC 74.82 ± 2.0), offered a favorable balance between accuracy and computational efficiency. Our findings demonstrate the potential of deep learning models in improving PCa lesion segmentation using CDI^s to enhance PCa management and clinical support.

In this work, we present VAIRO, a Vision-based Adaptive Impedance-control RObotic framework for the purpose of manipulating soft materials, centered around the use case of rolling croissant dough for use in artisanal bakeries. Traditional automated processes for the industrial production of croissants consist of overly bulky equipment and fail to preserve the artisanal quality of hand-rolled croissants, with one of the major challenges being the high variability in the dough properties. VAIRO addresses these challenges by introducing a novel vision-based adaptive Cartesian impedance control strategy for collaborative robot arms to regulate rolling forces in real-time without the need for estimating the properties of the soft material. As such, VAIRO mimics the tactile adjustments made by human pastry chefs, ensuring consistent layer thickness and eliminating gaps. Using a Kinova Gen3 robotic arm and a custom-designed end-effector, we demonstrate that VAIRO can successfully manipulate various "doughs" without estimating any material properties. These results are promising and offer a cost-effective, small-scale alternative for local craft bakeries to leverage automation while maintaining high artisanal quality.

Food intake monitoring is a crucial area of research in food computing due to its complexity and significant potential for improving health outcomes. While traditional 2D image-based dietary assessments provide basic information, video offers a more detailed understanding of both the quantity of food consumed and the manner in which it is eaten. However, current video-based dietary analysis remains limited to coarse metrics, such as counting bites. In this paper, we introduce FoodVideoQA, a novel approach that leverages Vision-Language Models (VLMs) to analyze food intake videos comprehensively. We discuss the inherent limitations of a VLM-based approach to this problem, demonstrating the necessity for further novel approaches in this field. This work paves the way for future studies for more advanced multimodal food intake measurement and behavioral studies. Source code is available at https://github.com/isobarbaric/FoodVideoQA.

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 developers working with SSL in ultrasound.

Food safety is essential for those who are vulnerable to foodborne illnesses. This study explores food degradation analysis using computer vision techniques combined with unsupervised machine learning. We extract visual features related to shape, texture, and color and apply fuzzy clustering to identify meaningful degradation states, capturing the gradual nature of food decay without rigid class boundaries. Additionally, we extract separate features using a Vision-Language Model (VLM) and integrate them into the clustering analysis. This multimodal approach enables both low-level visual feature analysis and high-level semantic interpretation of food degradation. Our study yields meaningful insights and lays the foundation for future research in food monitoring and safety.

Various weather modelling problems (e.g., weather forecasting, optimizing turbine placements, etc.) require ample access to high-resolution, highly accurate wind data. Acquiring such high-resolution wind data, however, remains a challenging and expensive endeavour. Traditional reconstruction approaches are typically either cost-effective or accurate, but not both. Deep learning methods, including diffusion models, have been proposed to resolve this trade-off by leveraging advances in natural image super-resolution. Wind data, however, is distinct from natural images, and wind super-resolvers often use upwards of 10 input channels, significantly more than the usual 3-channel RGB inputs in natural images. To better leverage a large number of conditioning variables in diffusion models, we present a generalization of classifier-free guidance (CFG) to multiple conditioning inputs. Our novel composite classifier-free guidance (CCFG) can be dropped into any pre-trained diffusion model trained with standard CFG dropout. We demonstrate that CCFG outputs are higher-fidelity than those from CFG on wind super-resolution tasks. We present WindDM, a diffusion model trained for industrial-scale wind dynamics reconstruction and leveraging CCFG. WindDM achieves state-of-the-art reconstruction quality among deep learning models and costs up to 1000× less than classical methods.

Current deep learning approaches for prostate cancer lesion segmentation achieve limited performance, with Dice scores of 0.32 or lower in large patient cohorts. To address this limitation, we investigate synthetic correlated diffusion imaging (CDI$^s$) as an enhancement to standard diffusion-based protocols. We conduct a comprehensive evaluation across six state-of-the-art segmentation architectures using 200 patients with co-registered CDI$^s$, diffusion-weighted imaging (DWI) and apparent diffusion coefficient (ADC) sequences. We demonstrate that CDI$^s$ integration reliably enhances or preserves segmentation performance in 94% of evaluated configurations, with individual architectures achieving up to 72.5% statistically significant relative improvement over baseline modalities. CDI$^s$ + DWI emerges as the safest enhancement pathway, achieving significant improvements in half of evaluated architectures with zero instances of degradation. Since CDI$^s$ derives from existing DWI acquisitions without requiring additional scan time or architectural modifications, it enables immediate deployment in clinical workflows. Our results establish validated integration pathways for CDI$^s$ as a practical drop-in enhancement for PCa lesion segmentation tasks across diverse deep learning architectures.

Breast cancer was diagnosed for over 7.8 million women between 2015 to 2020. Grading plays a vital role in breast cancer treatment planning. However, the current tumor grading method involves extracting tissue from patients, leading to stress, discomfort, and high medical costs. A recent paper leveraging volumetric deep radiomic features from synthetic correlated diffusion imaging (CDI^s) for breast cancer grade prediction showed immense promise for noninvasive methods for grading. Motivated by the impact of CDI^s optimization for prostate cancer delineation, this paper examines using optimized CDI^s to improve breast cancer grade prediction. We fuse the optimized CDI^s signal with diffusion-weighted imaging (DWI) to create a multiparametric MRI for each patient. Using a larger patient cohort and training across all the layers of a pretrained MONAI model, we achieve a leave-one-out cross-validation accuracy of 95.79 that previously reported.