Computer Vision news and analysis on Web Pulse: 3421 curated articles tracking the latest Computer Vision developments, tools, and research, updated continuously from vetted sources.
Researchers have developed Geometry-Aware Representation Denoising (GARD), a framework that uses diffusion-based restoration within a feed-forward 3D reconstruction model's feature space to improve multi-view 3D reconstr…
A new study benchmarked twelve deep learning architectures across four model families—convolutional neural networks, vision transformers, hybrid CNN-transformer backbones, and vision-language models—for multi-disease ret…
Researchers have developed DecoupleGen, a method that uses personalized text-to-image diffusion models to generate training images of objects in uncommon contexts, such as a beach ball on a road. The approach aims to cor…
Researchers have developed Evi-Steer, a new framework for BiomedCLIP that enables uncertainty-aware fine-tuning of biomedical vision-language models by updating only 0.11% of total parameters. The method uses evidential …
Researchers have developed a neural rule evaluator that detects anomalies by identifying violations of logical constraints, even when such violations are absent from training data. The system compiles rules into directed…
Human Archive has raised $8.2 million from Wing Venture Capital, NVP Capital, Y Combinator, and angel investors from OpenAI, Nvidia, Google, and Meta to deploy head-mounted and wrist-mounted cameras on workers in residen…
Pixal3D, a new image-to-3D AI model, focuses on preserving pixel-level alignment between a source image and the generated 3D asset, addressing common fidelity issues where models lose shape or detail. Developers can inte…
New Zealand Health Minister Simeon Brown announced the government has begun procurement and testing of an artificial intelligence tool to assist in reading breast screening mammograms, with a planned rollout from early 2…
Human Archive, a company backed by $8.2 million from investors including Wing Venture Capital and Y Combinator, is deploying head-mounted cameras on workers in India to collect first-person video data for training AI mod…
AI video firm Runway announced Project Luxo on Tuesday, a research initiative debuting three short films produced by a single staff member for roughly $4,000 each, aiming to demonstrate that AI-generated media can achiev…
Google has released Gemini Omni, a multimodal AI model designed to process and edit video natively within a single architecture rather than converting it into separate formats. The model can analyze full video clips for …
A developer built a tool that converts UI screenshots into live, rendered React and Tailwind CSS components in under three hours. The system uses a Go backend to compress and send images to Claude Vision, streaming the g…
A developer has released one of the first publicly documented LoRA training runs and general-purpose visual-enhancement LoRAs for HiDream-O1-Image, a top-ranked open-weight text-to-image model that shipped inference-only…
BusPatrol, a company that has installed AI-powered cameras on over 40,000 school buses across 24 states, plans to expand its surveillance system to record all passing license plates and share that data with law enforceme…
Xiangyu Chen is a Senior Research Engineer in the Autonomous Vehicle Research Group at NVIDIA, where he focuses on translating heterogeneous multimodal data into learning objectives for robust autonomous systems. Before …
Jason Stock, a researcher specializing in artificial intelligence, machine learning, computer vision, and generative AI, focuses his work on climate simulation. His contributions to the field are documented through his G…
Microsoft released Lens, a 3.8-billion-parameter foundational text-to-image diffusion model designed for efficient training and fast high-resolution generation. The model combines dense-caption pre-training, mixed-resolu…
Human Archive, a Silicon Valley startup founded by UC Berkeley and Stanford students, has raised $8.2 million to equip Indian gig economy workers with camera-equipped caps that collect first-person video data for trainin…
Google's Gemini models can analyze images, video, and documents without task-specific training, performing tasks like reading product labels, counting objects, and extracting fields from a well-written prompt. The models…
OpenAI released a new generation of vision-capable multimodal models, including GPT-5, GPT-5 Mini, GPT-5 Nano, and GPT-5.1, that can analyze images and perform visual reasoning directly from text prompts without task-spe…