fangchangma/self-supervised-depth-completion
A self-supervised deep learning model for estimating dense depth maps from sparse LiDAR data and monocular camera images.
Not currently ranked — collecting fresh signals.
star history
This repository implements a PyTorch-based neural network for self-supervised depth completion, trained on the KITTI dataset. The model takes sparse LiDAR depth measurements and RGB camera images as input and predicts full dense depth maps without requiring ground-truth annotations. It uses a self-supervised training approach based on view synthesis, where photometric consistency across frames serves as the supervision signal.
Frequently asked
- What is fangchangma/self-supervised-depth-completion?
- A self-supervised deep learning model for estimating dense depth maps from sparse LiDAR data and monocular camera images.
- Is self-supervised-depth-completion open source?
- Yes — fangchangma/self-supervised-depth-completion is open source, released under the MIT license.
- What language is self-supervised-depth-completion written in?
- fangchangma/self-supervised-depth-completion is primarily written in Python.
- How popular is self-supervised-depth-completion?
- fangchangma/self-supervised-depth-completion has 654 stars on GitHub.
- Where can I find self-supervised-depth-completion?
- fangchangma/self-supervised-depth-completion is on GitHub at https://github.com/fangchangma/self-supervised-depth-completion.