computer vision - culture, news & data
Joined August 2014
- Tweets3.5K
- Following5.8K
- Followers18.1K
- Likes66
The Official Stats of #ECCV2026 !!!
#ECCV2026 is just amazing! Around 6,000 to over 7,000 participants.
Submissions: 10,473 papers submitted.
Accepted Papers: 2,883 accepted papers. September 8–12, 2026. Malmö
Machine Learning, a free book :)
Download 348-page PDF eBook >> Machine Learning for Scientists and Engineers: smlbook.org/book/sml-book-dr…
————
#DataScience #DataScientist #ML #Mathematics #Algorithms
Book: Mathematical Methods in Data Science: Bridging Theory and Applications with Python, by S. Roch mmids-textbook.github.io
Google shipped Gemini 3.7 Flash: same 1M-token multimodal input (image, video, audio, PDF) as its predecessor, now sharper. #ComputerVision #Multimodal
9to5google.com/2026/08/13/ge…
Zoom out on August: nearly every major model release this month — Gemini, GLM, Qwen, Grok — ships with vision built in by default. Multimodal isn’t a feature anymore, it’s the baseline. #ComputerVision #Multimodal
Fresh off arXiv: YOLO26 — a unified real-time vision model with NMS-free inference, doing detection, segmentation AND pose estimation in one architecture. Real-time CV keeps getting faster and more general. #ComputerVision #YOLO
huggingface.co/papers/trendi…
Papers with Code #CVPR2026: Browse accepted papers from CVPR 2026 in Denver, grouped by research area and task, with code, project links, evals and Hugging Face artifacts where available. #ComputerVision paperswithcode.co/conference…
Best AI Models in August 2026: Updated Rankings and Comparisons - Best AI Models in 2026
Rankings, comparisons, and deep dives, updated monthly as new models ship. Claude Opus 5 still sits on top, but ... felloai.com/best-ai-models/
Baidu's Apollo Go launched its first European robotaxi trial in London on July 28: camera + LiDAR perception stacks now being road-tested outside their home markets. #ComputerVision #AutonomousDriving
autonomous-vehicles-conferen…
Google added AI image generation to Google Earth, letting anyone superimpose fake photorealistic scenes onto real satellite maps. Pulled within 24h after people generated fake disasters and explosions. cbc.ca/news/world/google-ear…
This month's clearest AI trend: vision-language-action models are leaving the browser tab behind and getting wired into real warehouse hardware at scale — an "Embodied AI" boom turning perception into physical action. #ComputerVision #EmbodiedAI
imfounder.com/science-tech/a…
Xpeng unveiled X-Mind at CVPR 2026: a world-model framework that lets self-driving cars run a spatial-temporal simulation of "what happens next" before generating a driving action, instead of just reacting. #ComputerVision #WorldModels
amenaauto.com/wheels-wings-a…
Quiet but huge: the FDA’s AI-enabled device list just crossed 1,524 cleared algorithms, 76% of them radiology — 68 new ones cleared in Q1 2026 alone. Medical imaging is where CV ships fastest. #ComputerVision #MedicalImaging
radiologybusiness.com/topics…
IJCB 2026 (Rome, Sept 1–4) is running a competition on face recognition trained only on synthetic data, no real faces. A real stress test for how far synthetic data can replace sensitive biometric datasets. #FaceRecognition #ComputerVision #IJCB2026
biometricupdate.com/202608/r…
Back after a long break!!! CV barely resembles the field from a few years ago: real-time segmentation, single-image 3D, agentic image reasoning and embodied action are all converging into one stack. Good time to start sharing again. 👋 #ComputerVision
4th workshop onVision-based InduStrialInspectiON (VISION)
VISION’26 @ CVPR: Vision-Based Industrial Inspection Across Modalities (Surface, X-ray, 3D & More)
vision-workshop-26.github.io…
Deepfake makers can now evade an unusual detection method | New Scientist newscientist.com/article/247…
