ant-research/CoDeF
A PyTorch implementation of CoDeF, a neural video representation using canonical content and temporal deformation fields to lift image algorithms to video processing.
Not currently ranked — collecting fresh signals.
star history
CoDeF proposes a video representation consisting of a canonical content field that aggregates static content across a video and a temporal deformation field that records transformations from the canonical image to each frame. These two fields are jointly optimized to reconstruct the target video through a tailored rendering pipeline. The approach enables lifting image-processing algorithms to video by applying them to the canonical image and propagating results via the deformation field.
Frequently asked
- What is ant-research/CoDeF?
- A PyTorch implementation of CoDeF, a neural video representation using canonical content and temporal deformation fields to lift image algorithms to video processing.
- Is CoDeF open source?
- Yes — ant-research/CoDeF is an open-source project tracked on heatdrop.
- What language is CoDeF written in?
- ant-research/CoDeF is primarily written in Python.
- How popular is CoDeF?
- ant-research/CoDeF has 4.9k stars on GitHub.
- Where can I find CoDeF?
- ant-research/CoDeF is on GitHub at https://github.com/ant-research/CoDeF.