AI repos for Computer vision
Open-source AI projects built around Computer vision or genuinely useful if you work with it, each with what it does for you. 10 repos, most relevant first.
-
lingbot-map
Reconstructs 3D point clouds and camera poses from video frames one at a time.
-
AI-For-Beginners
Has a dedicated computer vision track with hands-on notebooks and labs.
-
upscayl
Run Real-ESRGAN super-resolution models locally to enlarge and sharpen images, with swappable models and batch mode.
-
Unlimited-OCR
Vision-language model that parses document images and PDFs into structured text.
-
glance-speedlab
Benchmark harness with per-stage latency reporting for live-camera vision-language inference, including logged failed experiments.
-
transformers
Pretrained vision models available through the same pipeline API.
-
ai-engineering-from-scratch
Includes a computer vision phase within the 20-phase curriculum of hands-on lessons.
-
InvokeAI
Generate, inpaint and refine images locally with diffusion models in a node-based editor.
-
screenpipe
Event-driven screen capture with OCR makes everything shown on screen searchable.
-
openshorts
Uses face-tracked, layout-aware cropping to turn landscape video into 9:16 clips.