Computer Vision news and analysis on Web Pulse: 4296 curated articles tracking the latest Computer Vision developments, tools, and research, updated continuously from vetted sources.
A developer built LankaLens, a mobile app that uses an open-weight AI model running locally to identify plants, trees, insects and other wildlife from photos taken outdoors. The app returns a short explanation and an obs…
Scale AI and Elorian released the Humanity's Sixth Sense (HSS) visual-reasoning benchmark on October 7th, 2026, in which the top model tested, OpenAI's GPT-6-astra, answered 53.6% of tasks correctly against 93.1% for 20 …
Researchers proposed MoR-MLLM, a computation-sparse multimodal large language model built on the Mixture-of-Recursions (MoR) framework that uses adaptive per-token recursion to dynamically adjust recursive depth, allocat…
Researchers established a comprehensive benchmark for X-ray radiology report generation on the CheXpert Plus dataset, which was released without baseline implementations or evaluation results, and proposed a new framewor…
A new method called SlideRuler calibrates pathology foundation models from a single slide scan, reducing mean target-to-source embedding distance by 16.3-38.5% relative to raw embeddings across two encoders and five SCOR…
Researchers proposed RACER, a training-free reflective agentic framework that decomposes long-video frame selection into query interpretation by a lightweight video large language model and evidence localization by an em…
Researchers introduced S2Tok, a feed-forward framework that maintains a size-adaptive, persistent scene state from uncalibrated image streams to enable streaming 3D Gaussian reconstruction, according to arXiv paper 2610.…
Researchers introduced DISRQAD, a subjective-quality dataset and diagnostic benchmark containing mean opinion scores for 14,000 super-resolution outputs from ten diffusion and four non-diffusion methods across four low-r…
A new arXiv paper (2610.09031v1) introduces a plug-and-play framework for concept-based interpretation and refinement of medical imaging models, building a Concept Bottleneck Model by aligning a single-modality encoder t…
Researchers introduced GUARD, a geometric uncertainty-aware framework that performs point filtering and segmentation in a single forward pass for robotic hard disk drive (HDD) disassembly, improving PointNet++ segmentati…
Microsoft researchers released Anaximander, an open-source system that runs geospatial deep learning models on any compute backend behind one interactive interface, with code at github.com/microsoft/nxmndr. The system pa…
SPLATIFY, a multi-agent framework described in arXiv paper 2610.09116v1, converts 3D Gaussian Splatting (3DGS) papers into trainable gsplat-based implementations where generic paper-to-code methods and frontier models fa…
A developer outlined a practical workflow for using AI inpainting to remove logos, timestamps, and text overlays from demo videos, emphasizing frame-by-frame consistency to avoid flicker. The guide recommends testing on …
Mistral AI has deprecated Pixtral 12B, the text-and-image model released on 2024-09-01 with a 128k context window and 128k max output, according to a provider announcement tracked by models.dev. No retirement date has be…
Apple researchers Jiatao Gu, Tianrong Chen, Ying Shen, David Berthelot, Shuangfei Zhai and Josh Susskind published Normalizing Trajectory Models (NTM) in October 2026, a generative modeling method that treats each revers…
A developer used Hugging Face's ML Intern agent to build a 0.8B prompt-rewriter model that runs on a CPU and returns valid output 99.7% of the time, after the official Qwen-Image 2.1 prompt rewriter required a 9B model a…
Liquid AI's d1 decision model is now available on Vercel's AI Gateway, where it evaluates shared state against typed questions for classification, routing, and scoring and returns structured answers with probabilities wi…
Lichess added image import to its Board Editor on October 7, 2026, letting users select "From image" or drag and drop a picture of a board position onto the board. The feature runs on the ChessQ Lite V4 recognition model…
Researchers introduced DeCoPrune, a KV-cache pruning method for autoregressive video diffusion that uses denoising consistency to compress the continuously growing cache. The work targets the memory cost of streaming vid…
Liquid AI released d1-3B, an open-weight text-and-image decision model scoring an average 74.1 across 11 public image benchmarks, and the experimental d1-omni-600M for text paired with images or audio, on October 7th via…