shariqfarooq123/AdaBins
A deep learning model that estimates depth from single images by dividing the depth range into adaptive bins.
Not currently ranked — collecting fresh signals.
star history
AdaBins is an official implementation of a depth estimation method that uses a U-Net architecture with adaptive binning to predict metric depth from monocular images. The model divides the depth range into learnable bins whose centers adapt based on the input image content, and regresses bin edges alongside depth values. Pretrained weights are provided for NYU-Depth-v2 and KITTI datasets.
Frequently asked
- What is shariqfarooq123/AdaBins?
- A deep learning model that estimates depth from single images by dividing the depth range into adaptive bins.
- Is AdaBins open source?
- Yes — shariqfarooq123/AdaBins is open source, released under the GPL-3.0 license.
- What language is AdaBins written in?
- shariqfarooq123/AdaBins is primarily written in Python.
- How popular is AdaBins?
- shariqfarooq123/AdaBins has 785 stars on GitHub.
- Where can I find AdaBins?
- shariqfarooq123/AdaBins is on GitHub at https://github.com/shariqfarooq123/AdaBins.