astra-vision/MonoScene

[CVPR 2022] "MonoScene: Monocular 3D Semantic Scene Completion": 3D Semantic Occupancy Prediction from a single image

44
/ 100
Emerging

This project helps self-driving car engineers and robotics researchers understand and interact with 3D environments from a single camera image. It takes a standard 2D image and generates a complete 3D scene, detailing the occupancy and semantic labels (e.g., road, car, building) for every voxel in that space. This allows for a rich 3D perception of the surroundings, which is crucial for navigation and object interaction.

799 stars. No commits in the last 6 months.

Use this if you need to reconstruct a detailed 3D environment, including identifying objects and their spatial relationships, using only monocular camera input from urban or indoor scenes.

Not ideal if your application requires real-time processing on embedded systems with limited computational resources, or if you need extremely fine-grained object detection rather than scene-level understanding.

autonomous-driving robotics-perception 3d-reconstruction scene-understanding computer-vision
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 18 / 25

How are scores calculated?

Stars

799

Forks

75

Language

Python

License

Apache-2.0

Last pushed

Apr 06, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/computer-vision/astra-vision/MonoScene"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.