Raessan/dinov3_deepstream
DeepStream integration of Meta’s DINOv3 backbone with lightweight heads for vision tasks.
This application helps operations engineers and robotics developers process live video feeds from cameras, files, or streams in real time. It takes raw video input and simultaneously produces outputs like identified objects (with bounding boxes), semantic segmentation masks (pixel-level classification), depth maps, and optical flow vectors, all at maximum speed and efficiency. This is ideal for scenarios requiring simultaneous, low-latency analysis of visual data.
Use this if you need to run multiple real-time vision tasks like object detection, depth estimation, and segmentation on video streams from cameras or files using NVIDIA GPUs or Jetson devices for optimal performance.
Not ideal if you primarily need to perform a single vision task, do not have access to NVIDIA hardware, or are working with still images rather than continuous video streams.
Stars
19
Forks
—
Language
C++
License
Apache-2.0
Category
Last pushed
Feb 05, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/computer-vision/Raessan/dinov3_deepstream"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
isl-org/Open3D
Open3D: A Modern Library for 3D Data Processing
cvg/Hierarchical-Localization
Visual localization made easy with hloc
gmberton/CosPlace
Official code for CVPR 2022 paper "Rethinking Visual Geo-localization for Large-Scale Applications"
Vincentqyw/image-matching-webui
🤗 image matching webui
cvg/glue-factory
Training library for local feature detection and matching