Computer Vision news and analysis on Web Pulse: 3326 curated articles tracking the latest Computer Vision developments, tools, and research, updated continuously from vetted sources.
Researchers present a training algorithm that reduces optimization variance in small neural networks used for real-time neural shading, improving convergence reliability. The method uses an ensemble of network instances,…
Apple Inc. is aiming to launch camera-equipped AirPods this fall, according to Bloomberg's Mark Gurman, with the design codenamed B790 already referenced in the OS 27 betas. The cameras, which see both standard and infra…
Roboflow published a tutorial demonstrating human-object interaction (HOI) detection without training an interaction model, using RF-DETR to detect workers, forklifts, pallets, and carts (77.5% mAP@50) and Gemini 2.5 Pro…
A developer building claude-real-video, an MIT-licensed local pipeline for feeding videos to vision LLMs, found that frame selection is the critical factor in video understanding, as LLMs can afford about 150 images per …
Knox County Mayor Glenn Jacobs, a former WWE star known as Kane, has paused a $1 million deal to install 270 Flock Safety cameras across the county, citing concerns that the AI-powered surveillance technology violates Fo…
A study by computer scientists at Bradford University using facial recognition to analyze Holbein's drawings concluded that a drawing inscribed 'Anna Bollein Queen' was not Anne Boleyn, but art historian Bendor Grosvenor…
A developer detailed the performance overhaul of AURA, a desktop AR companion that reads facial expressions and hand gestures. The project reduced its cold launch time from about 35 seconds to 1.5 seconds by lazy-loading…
The Guardian's Football Weekly podcast revealed that artificial intelligence is reshaping football scouting, transfer gossip, and tactical analysis, with clubs increasingly relying on machine learning models for recruitm…
A college student built FocusFleet, a drowsy driver detection system that uses MediaPipe to track 468 facial landmarks and a custom-trained Keras model to monitor eye closure and yawning, triggering an audio alert when i…
A developer has created a universal field study skill that converts any reference photo into a minimalist editorial illustration. The skill, available as a SKILL.md file, uses AI to identify the subject type and generate…
Researchers introduce WaiT, a Wavelet-aware image Transformer that decomposes image generation into coarse and fine bands via lossless wavelets, allowing high-frequency bands to remain pure noise until coarse structure e…
Researchers proposed an automated classification method for COVID-19 cases from chest CT volumes using a 2D/3D hybrid convolutional neural network (CNN) designed for anisotropic volumes. The method achieved an average cl…
Researchers at an undisclosed institution present CV, a computer vision framework that estimates the fatigue life of lightweight alloy steels from optical micrographs, achieving an R² of 0.93 and RMSE of 0.18 log-cycles …
A new arXiv preprint (2607.28868v1) introduces a physics-aligned augmentation procedure for self-supervised learning in scientific imaging, tested on electron microscopy and 4D-STEM diffraction across five SSL paradigms …
Researchers propose SCMA, a structure-conditioned and metal-aware Flow Matching framework for CT metal artifact reduction, which outperforms representative methods in suppressing artifacts and preserving anatomy on simul…
Researchers propose mirror learning, a framework that trains AI policies from third-person observation by using a fine-tuned video diffusion model to transform perspectives and an inverse dynamics model to infer actions,…
Researchers introduced SERUM, a multi-pass framework that extracts finite-state behavioral models from unstructured egocentric video using hierarchical VLM annotation, achieving lower perplexity and higher action predict…
A new benchmark study from arXiv compares five deep learning models for 3D brain tumor segmentation from MRI, finding trade-offs between accuracy and efficiency. The study evaluated 3D U-Net, SegResNet, Swin UNETR, SegMa…
A new inference-time framework, Adjudicated Captioning, achieves CIDEr 117.6 and SPICE 21.9 on the COCO Karpathy test split for zero-shot image captioning, up from 108.0 and 20.3 for the baseline IFCap, a +9.6 CIDEr gain…
Researchers from Southeast University propose RAID, a new AI-generated image detection method that uses bit-reversed images and a gradient-based patch selection with a convolutional classifier, achieving state-of-the-art…