twhui/LiteFlowNet
A lightweight convolutional neural network for optical flow estimation in video sequences.
Not currently ranked — collecting fresh signals.
star history
LiteFlowNet is a computer vision model that estimates optical flow from image pairs. It employs a pyramidal feature extraction, cascaded flow inference with cost volumes and sub-pixel refinement, feature warping layers, and feature-driven local convolution for flow regularization. The model targets real-time performance with improved accuracy over comparable networks like PWC-Net on benchmarks like KITTI.
Frequently asked
- What is twhui/LiteFlowNet?
- A lightweight convolutional neural network for optical flow estimation in video sequences.
- Is LiteFlowNet open source?
- Yes — twhui/LiteFlowNet is an open-source project tracked on heatdrop.
- What language is LiteFlowNet written in?
- twhui/LiteFlowNet is primarily written in C++.
- How popular is LiteFlowNet?
- twhui/LiteFlowNet has 633 stars on GitHub.
- Where can I find LiteFlowNet?
- twhui/LiteFlowNet is on GitHub at https://github.com/twhui/LiteFlowNet.