Computer Vision news and analysis on Web Pulse: 1698 curated articles tracking the latest Computer Vision developments, tools, and research, updated continuously from vetted sources.
Bytedance Seedance V1.5 Pro Image To Video defeated Marey Realism V1.5 in a head-to-head test, scoring 17.2 to 13.4, because it better obeyed the prompt's instructions for specific physical events and scene geography. Th…
China's CPU-only LineShine supercomputer, built on the Arm architecture, claimed the top spot on the 67th TOP500 list, achieving 2.198 Exaflops of sustained FP64 performance with over 13 million cores. The system challen…
Rundiffusion Photo Flux defeated Bagel 20.4 to 11.8 in a head-to-head image generation test, winning on instruction following and detail accuracy despite Bagel producing a more aesthetically convincing single image. The …
Russian Zala drones equipped with Nvidia TX2-A AI chips are attacking Ukraine, using onboard artificial intelligence to recognize and target vehicles, people, and military markings. The AI allows the drones to pre-sort t…
Developer Priyanka built an AI-powered app that analyzes desktop screenshots using NVIDIA NIM's vision model and prompt engineering to generate roasts, productivity advice, or interview-style feedback. The project demons…
Developer tryfonaskam released PILA, an open-source imitation learning agent that learns to play the browser racing game PolyTrack by training on screen captures and human keyboard inputs. Implemented in PyTorch, the pip…
Researchers have developed a new type of pixel that can both emit and receive light, potentially enabling more efficient and compact display and sensor technologies. The innovation could lead to advancements in augmented…
A developer building the V.E.L.O.C.I.T.Y.-OS bare-metal kernel implemented a double-buffered graphical user interface called Synaptic Canvas GUI, running on UEFI GOP framebuffer without floating-point libraries. The GUI …
9 Mothers, a Y Combinator-backed defense startup, is selling Edda, an AI-powered point-defense turret designed to stop small FPV suicide drones at close range. The Austin-based company has sold $1.6 million of systems to…
A developer launched FencePreview, an AI-powered tool that generates photorealistic fence previews from yard photos, enabling contractors to close deals on-site by showing customers exactly how their new fence will look.
Flock Safety's AI-powered surveillance cameras, which track vehicles and individuals beyond just license plates, have rapidly proliferated to over 100,000 units nationwide, raising privacy and civil liberties concerns du…
A developer built a hands-free computer interface using a webcam and microphone, combining head tracking with MediaPipe FaceMesh and voice commands. The system enables mouse control via head gestures, clicking with eye b…
A developer built PSLRate, a browser-only face-rating app that computes attractiveness scores entirely client-side using MediaPipe's FaceLandmarker, with no image uploads or per-scan costs. The free tier analyzes facial …
A developer built a production-grade cross-platform face recognition pipeline with anti-spoofing, using a lightweight RFB-320 face detector, FaceNet-style embeddings, and HNSW indexing for sub-millisecond matching. The s…
Drivers in the UK face six penalty points for using mobile phones while driving, a law that is frequently broken. New AI cameras are being trialed in Scotland to catch offenders, with potential nationwide rollout. Expert…
Full Stack Developer & AI/ML Specialist Rahul Agarwal built GitPulse, a cinematic web application that transforms standard GitHub profiles into animated contribution universes. The app uses Canvas API, GSAP, and Vanilla …
Hugging Face's top-voted AI papers for June 28, 2026, reveal a shift from models that 'know how to answer' to systems that 'know how to act, remember, adapt, and generate context-aware content.' The roundup covers 10 not…
Artist Trevor Paglen's new book 'How to See Like a Machine: Images After AI' examines how AI-generated imagery is replacing human-made visual art, arguing that algorithms trained to capture attention are eroding the inte…
AI training startup Shift offers free housecleaning in exchange for customers allowing a camera headset on the cleaner to record footage, which is then licensed to train AI-powered household robots. The company, owned by…
NVIDIA released LocateAnything-3B, a vision-language model designed for visual grounding that can identify and locate objects in complex scenes using natural language queries. The 3-billion-parameter model, trained on 7 …