akirasosa/mobile-semantic-segmentation
A semantic segmentation model using MobileNetV2-inspired U-Net architecture for real-time mobile hair segmentation.
Not currently ranked — collecting fresh signals.
star history
This project implements a deep learning model for real-time semantic segmentation optimized for mobile deployment. The architecture combines MobileNetV2 depthwise separable convolutions with a U-Net encoder-decoder structure to achieve 0.89 IoU on face/hair segmentation. It includes a training pipeline with dice coefficient loss, supports both PyTorch and TensorFlow, and exports to CoreML for iOS deployment and TensorFlow Lite for Android.
Frequently asked
- What is akirasosa/mobile-semantic-segmentation?
- A semantic segmentation model using MobileNetV2-inspired U-Net architecture for real-time mobile hair segmentation.
- Is mobile-semantic-segmentation open source?
- Yes — akirasosa/mobile-semantic-segmentation is open source, released under the MIT license.
- What language is mobile-semantic-segmentation written in?
- akirasosa/mobile-semantic-segmentation is primarily written in Python.
- How popular is mobile-semantic-segmentation?
- akirasosa/mobile-semantic-segmentation has 721 stars on GitHub.
- Where can I find mobile-semantic-segmentation?
- akirasosa/mobile-semantic-segmentation is on GitHub at https://github.com/akirasosa/mobile-semantic-segmentation.