Segmentation Pytorch Save

Semantic Segmentation Pytorch

    author is leilei
    Restart this project from 2017-10-01 
    
    Now the 1.alpha.0 version has been basically completed, to be tested.
    TODO
    Add distributed and optimize code.

Environment

    python: 3.6+
    ubuntu16.04 or 18.04
    pytorch 1.6 (cuda10.2 docker)
    tensorboard 2.0
    scikit-learn 0.24.1

Note

• If a black border is introduced, it will be regarded as one type, and the default is 0 !
• label value is [1, N], 0 is black border class !
• Not supporting distributed(NCCL), just support DataParallel.

Getting Started

• How to Use

Demo

• Just see demo.py

Evaluation index

• Just see metrics.py
• Support acc, mean_precision, mean_recall, mean_iou

Support Network

• deeplab_v3_plus
• pspnet
• unet
• spp-net
• HF_FCN
• deeplab_v3
• HRNet
• U^2Net
• ...

Data Aug

• data-augumentations

support 
    random zoom-in/out, random noise,
    random blur, random color-jitter(brightness-contrast-saturation-hue)
    random affine, random rotate, random flip

Others

• building-segmentation-dataset
• reflective-clothes-detect-dataset