Image recognition, detection, and visual AI
![[2602.09678] Administrative Law's Fourth Settlement: AI and the Capability-Accountability Trap](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
Abstract page for arXiv paper 2602.09678: Administrative Law's Fourth Settlement: AI and the Capability-Accountability Trap
Abstract page for arXiv paper 2601.13622: CARPE: Context-Aware Image Representation Prioritization via Ensemble for Large Vision-Language...
Abstract page for arXiv paper 2603.26551: Beyond MACs: Hardware Efficient Architecture Design for Vision Backbones
The paper presents Texo, a compact formula recognition model with 20 million parameters, achieving high performance comparable to larger ...
The paper introduces Bonsai, a framework for accelerating Convolutional Neural Networks (CNNs) through criterion-based pruning, demonstra...
The paper explores how narrow fine-tuning of vision-language agents can lead to significant safety alignment issues, highlighting the ris...
The AIdentifyAGE ontology aims to enhance forensic dental age assessment by providing a standardized framework for integrating clinical, ...
The article discusses the implications of Google's Pomelli feature, which generates product visuals using AI, raising questions about cre...
Makimus-AI is a free, open-source local app that enables users to search their image libraries using natural language queries, functionin...
This Reddit thread serves as a community hub for discussions and updates regarding the decisions for CVPR‘26, a prominent conference in c...
The Qwen3.5 model trains on visual-text tokens natively, potentially addressing the 'modality gap' found in CLIP-based models, enhancing ...
The article discusses user frustrations with the recent performance issues of GPT-5.2, highlighting problems with OCR accuracy and file g...
This article presents a novel imaging algorithm that utilizes strong scattering to achieve super-resolution in dynamic random media, enha...
This paper introduces View Invariant Learning (VIL) for enhancing Vision-Language Navigation in Continuous Environments (VLNCE), addressi...
The paper presents Filter2Noise, a novel framework for interpretable and zero-shot low-dose CT image denoising, achieving state-of-the-ar...
VIRENA is a novel platform designed for controlled experimentation in social media environments, enabling researchers to study human-AI i...
This article presents a novel demand estimation method that utilizes unstructured data from text and images to enhance substitution patte...
This paper explores the integration of vision-language models in autonomous driving, focusing on safety assessment and decision-making th...
The paper presents LMSeg, a novel approach for open-vocabulary semantic segmentation that enhances visual and linguistic feature alignmen...
This article examines whether vision-language models (VLMs) respect contextual integrity when disclosing location information, highlighti...
This article presents a novel approach to medical imaging classification using autoassociative learning, demonstrating improved accuracy ...
This paper presents USplat4D, a novel framework for monocular 4D reconstruction that incorporates uncertainty in dynamic Gaussian splatti...
The paper presents AliAd, a model for multimodal multiview human activity recognition that enhances performance by integrating diverse vi...
Get the latest news, tools, and insights delivered to your inbox.
Daily or weekly digest • Unsubscribe anytime