Computer Vision news and analysis on Web Pulse: 4324 curated articles tracking the latest Computer Vision developments, tools, and research, updated continuously from vetted sources.
Researchers introduced LLaDA-UI, a block-wise diffusion vision-language model designed for GUI agents that must repeatedly perceive screen states and emit actions. The work applies diffusion large language models' block-…
A developer has published a tutorial for building "Posture Guardian," a browser-based tool that uses MediaPipe's pose estimation model and Vue.js to detect slouching, forward head posture, and uneven shoulders in real ti…
A developer documented building a small handwriting classifier, shipping it in a Flutter app, and teaching it new characters from a few local samples. The same collection of posts covers cross-compiling Obscura and V8 fo…
An independent research group in Austria has developed two quantization techniques, GSQ and RCO, that compress Z.ai's 320-billion-parameter GLM 5.3 Flash vision-language model from roughly 320GB at full precision to unde…
Muon Space posted an on-site software engineer role in San Jose, CA for computer vision and machine learning at a salary range of $212,000 to $233,000 per year, plus equity and benefits. The listing, filed under Inferenc…
1X Technologies posted an on-site AI Researcher - Perception role in San Carlos, CA, paying $200k–300k/yr, which sits 5% above the $237k median for Core ML roles on the board that publish pay (516 of 663). The hire will …
Researchers introduced Parallel Decoding Distillation (PDD), a trajectory-based distillation method that accelerates image and video generation by predicting multiple denoising steps per network evaluation, achieving sta…
Research papers including FilmAgent, VideoGen-of-Thought, CineAGI, StoryAgent, and AniME are decomposing video production into specialized agents coordinated by an orchestrator, with CineAGI reporting meaningful gains in…
President Donald Trump said Sunday aboard Air Force One that he likes Flock Safety's AI-powered surveillance cameras because they help law enforcement, his clearest public stance yet on the technology. Trump acknowledged…
Skyvern released an AI browser automation tool that uses computer vision and large language models to fill web forms from natural-language task descriptions, according to the company's developer documentation. Skyvern sa…
IBM and NASA released the NASA-IBM Lunar Foundation Model as open source on Hugging Face under the Prithvi family, cutting root-mean-square error in flagging likely lunar ice deposits by up to 22% against a SwinV2-B base…
A developer released neural_texture_es2, a self-contained C++ experiment that encodes an RGB image or up to four PBR material textures as a shared low-resolution latent texture plus a tiny MLP decoder, trained entirely w…
Researchers Florian Hahlbohm, Jorge Condor, Linus Franke, Martin Eisemann and Marcus Magnor published an arXiv paper introducing a compact neural appearance model for 3D Gaussian Splatting that reduces the per-primitive …
Pinterest expanded its partnership with Nvidia, pairing Nvidia's Blackwell GPUs and Dynamo inference framework with Pinterest's proprietary visual embeddings to achieve an 85x improvement in response startup times and a …
Adversarial fashion uses specific patterns and silhouettes to lower the confidence scores of object detection models or force them to misclassify humans as animals or objects, according to reporting on the practice. In 2…
A single 2B-parameter vision-language model placed first in the ≤2B parameter division of the EgoLongQA track of the Wearable-AI Challenge at ECCV 2026, scoring 0.8279 on the held-out test set. The system, described by i…
A submission to the EgoProactive track of the ECCV 2026 Wearable AI Challenge ranked first in the large-model division and second in the <=2B division, according to the team behind the entry. The task requires a wearable…
Parents and children in Illinois and California filed a proposed class action lawsuit against Meta over its dormant "NameTag" facial recognition software for smart glasses, alleging the underlying faceprint data could be…
A new method called TRACE (Trajectory-robust Admission with Evidence Ordering) addresses the growing inference latency and memory costs of GUI agents that accumulate high-resolution screenshots as their trajectories unfo…
Nigerian defence-tech startup Terra Industries has raised an additional $18 million, bringing its seed funding to $52 million, to expand manufacturing and international deployment of its autonomous security systems. The …