Computer Vision news and analysis on Web Pulse: 4312 curated articles tracking the latest Computer Vision developments, tools, and research, updated continuously from vetted sources.
Researchers introduced MRFFU-Net, a deep learning architecture that adds a Multi-Resolution Feature Fusion (MRFF) module to U-Net-like encoder-decoder structures for MRI segmentation, according to an arXiv paper (arXiv:2…
A new arXiv paper (arXiv:2610.00294v1) introduces LENS-GRF, a permutation-invariant Lesion Evidence Network with Gated Residual Fusion for four-class acne severity grading. On the ACNE04 dataset, fully automated LENS-GRF…
A new arXiv paper (2610.00302v1) introduces GRDisaster, a multi-task geospatial reasoning framework that uses vision-language models to geolocalize and interpret crowdsourced disaster imagery. Built on a benchmark of 26,…
A reverse-engineering effort using Anthropic's Claude model and a real PSP connected over USB ran more than 170 hardware experiments to document the PSP's Graphics Engine (GE), after which PPSSPP's software renderer repr…
Roboflow published a guide detailing how to build a two-stage computer vision model cascade in Roboflow Workflows, in which RF-DETR runs on every tile and accepts detections at or above 0.70 confidence while uncertain de…
Tavus introduced Griffin, which the company calls the first "Human Interaction Model," an AI that conducts real-time video calls by processing facial expressions, tone of voice, and gestures. In a Tavus study, 48 percent…
NVIDIA released DIN Deploy, an open-source collection of C++ samples that pairs ONNX Runtime with the NVIDIA TensorRT RTX execution provider to run AI models locally on Windows and Linux. The repository covers OpenAI Whi…
Ideogram released Ideogram 4.5, a model it says can edit targeted areas of an image while leaving the rest intact, with native 2K resolution starting at 0.8 cents per image. Runway, Pika, and Leonardo AI are already part…
Researchers at the University of California, Los Angeles built an optical-neural processor that detects deepfake videos with an average accuracy of 97.79% while analyzing 15 Celeb-DF videos simultaneously in a single opt…
A developer published a production-oriented Flutter architecture for streaming camera frames from Meta smart glasses and other robotics/AI hardware, using a latest-frame bounded-buffer strategy to avoid stale frames and …
A solo developer published OpenDLSS-NR, an open-source Vulkan reimplementation of NVIDIA's DLSS 5 neural rendering network that the developer says matches the original byte-for-byte and runs on RTX 40 hardware NVIDIA loc…
Researchers at the Weizmann Institute of Science in Rehovot, Israel, led by Michal Irani, built an AI "brain decoder" that reconstructs what a person is looking at from high-resolution fMRI scans and can also predict bra…
TrackFish3D, a geometry-driven self-supervised framework for dense multi-camera 3D tracking of schooling fish, improves 3D Multi-Object Tracking Accuracy from 87.7% for the strongest baseline to 95.8% on its benchmark, a…
LoopVL, a recurrent vision-language model that combines Module-Loop and Model-Loop computation to iteratively update a unified vision-language state through shared modules, outperforms similarly sized and larger non-recu…
Researchers introduced modal kinetic typography, a method that animates a vector glyph by building motion from the glyph's natural vibration modes, computed via a finite-element eigenproblem assembled from the vector out…
A training-free framework called Inductive Visual Logic (IVL) achieves the highest aggregate accuracy across multiple distant out-of-distribution (OOD) benchmarks under two vision-language model backbones, according to a…
Team MSU placed third in the ACM MM 2026 GenText-Forensics challenge with a decomposed chain-of-thought pipeline that pairs a document tampering detector with two Qwen3-VL-32B vision-language models, each LoRA-adapted to…
Researchers introduced MOBA-VL, a 9B-parameter vision-language model trained with event-localized multi-turn reinforcement learning on game telemetry to deliver real-time commentary for Multiplayer Online Battle Arena es…
A developer built ClearPix, a static, backend-free watermark remover that runs ONNX Runtime Web inference entirely in the browser via WebGPU with a WASM fallback, so user images and videos are never uploaded. The enginee…
Researchers at the University of Alabama trained a model only on video footage captured prior to violent events — combining body motion, facial disposition and audio — to test whether trainable 'pre-violent' characterist…