Star-Clouds/CenterFace
CenterFace is an anchor-free face detection and alignment model optimized for edge device inference using the NCNN framework.
Not currently ranked — collecting fresh signals.
star history
CenterFace is a practical face detection and facial landmark alignment system designed for edge devices, with a model size of 7.3MB. It uses an anchor-free approach and achieves competitive results on the WIDER FACE and FDDB benchmark datasets. The implementation uses the NCNN neural network inference framework for efficient deployment on resource-constrained devices, with multi-scale and flip evaluation for improved accuracy.
Frequently asked
- What is Star-Clouds/CenterFace?
- CenterFace is an anchor-free face detection and alignment model optimized for edge device inference using the NCNN framework.
- Is CenterFace open source?
- Yes — Star-Clouds/CenterFace is open source, released under the MIT license.
- What language is CenterFace written in?
- Star-Clouds/CenterFace is primarily written in C++.
- How popular is CenterFace?
- Star-Clouds/CenterFace has 1.3k stars on GitHub.
- Where can I find CenterFace?
- Star-Clouds/CenterFace is on GitHub at https://github.com/Star-Clouds/CenterFace.