Computer Vision news and analysis on Web Pulse: 3312 curated articles tracking the latest Computer Vision developments, tools, and research, updated continuously from vetted sources.
A researcher replaced the standard scaled dot-product attention in vision transformers with SSOG (Separable Sum of Gaussians), a content-blind geometric attention mechanism that uses a few Gaussian atoms per head and a c…
Plantum's AI-powered plant identifier app for iOS is available for a lifetime subscription at $19.99, down from $59.99, through a StackSocial deal. The app identifies plants from tens of thousands of species, diagnoses d…
ReClaim, a computer-vision-powered lost-and-found system for campus communities, uses CLIP embeddings and cosine similarity to match lost items with found items. The system, built with FastAPI and PyTorch, enforces a 3:4…
McAfee research shows AI models can identify locations from holiday photos posted on Instagram or Facebook, enabling scams that steal bank details. In tests of more than 21,000 travel images, one model correctly identifi…
PawPrint, a community-powered lost-dog search tool, uses Google Gemini to rank community sighting photos by visual similarity to the missing dog's photo, helping owners prioritize clues while leaving the final decision t…
EmpowerAgro, an AI-assisted agronomic platform built by developers using Google's AI ecosystem, brings real-time crop disease detection to smallholder farmers in India. The platform combines on-device computer vision wit…
Fieldwire engineer Gaston Dombiak spent a weekend using Anthropic's Claude Code to teach himself the SLAM algorithm, with the AI writing a 14-lesson curriculum, generating synthetic test data, and producing a working pro…
An engineer detailed the development of Veridian Resonance, a 28-page digital comic generated using a hybrid GenAI production pipeline. The project involved engineering a strict 'Cel-Shaded' visual formula and using char…
Meta has filed a patent application with the US Patent and Trademark Office for a system that uses AI facial recognition with its smart glasses to identify people and create media files, according to a 44-page filing sub…
A developer built PawSense AI, a full-stack application that uses Google Gemini's multimodal vision to generate detailed canine profiles from photos, including breed analysis, personality archetypes, and a conversational…
A man in India used ChatGPT to locate a single lost clog among more than 1,000 pairs of footwear, a feat captured on video and shared on Instagram by user Shubhang Borkar, drawing over 2 million views. The viral clip ill…
Meta Platforms Technologies, the entity holding Meta's R&D and hardware intellectual property, published a patent for a camera system that uses facial recognition to identify people and log their actions, bundling clips …
Moonshot AI's benchmark reveals that most frontier large language models, including GPT-4o and GPT-5.6 Sol, fail to exceed 60% accuracy on visual perception tasks, highlighting a significant gap between reasoning and per…
Meta is patenting AI glasses that can identify people in real time, integrating an onboard camera, a neural processing unit, and a cloud-based identity graph to provide users with a 'social cheat sheet' of names and shar…
Flock Safety CEO Garrett Langley, facing backlash over the company's 120,000 license plate reading cameras across 49 states, said in a Bloomberg interview that 'we're not Big Brother' and announced platform changes inclu…
A developer built a zero-GPU-cost system that generates 30-minute ASMR rain videos daily from a single still image using Python physics simulations and ffmpeg's displace filter, cutting generation time from 7-8 minutes t…
A cybersecurity and computer science engineering student participating in Gen AI Academy's APAC Edition is building a 3D LiDAR-based system to assess shelter damage after natural disasters, alongside an AI platform for e…
A developer presented a secure and efficient image encryption scheme based on chaotic systems, highlighting the use of chaotic maps for pixel permutation and diffusion. The approach combines sensitivity to initial condit…
A 14-year-old student from Sri Lanka built FetchAI, a cyberpunk-styled web app that uses Google's Gemini API to identify dog breeds from uploaded photos and generate personalized short stories. The app, developed with va…
A smartphone's AI-enhanced camera failed to capture a solar eclipse, producing a distorted image, highlighting the limitations of computational photography in extreme lighting conditions. The incident, reported by Creati…