Implementation is not original papers. The purpose of this project is to get started with semantic segmentation and master the basic process.
FCN32/8、SegNet、U-Net Model published. Thank you for your support.
Item | Value | Item | Value |
---|---|---|---|
keras | 2.2.4 | OS | win10 |
tensorflow-gpu | 1.10/1.12 | Python | 3.6.7 |
https://github.com/divamgupta/image-segmentation-keras
https://github.com/ykamikawa/SegNet
Data:
https://drive.google.com/file/d/0B0d9ZiqAgFkiOHR1NTJhWVJMNEU/view?usp=sharing
Data or:
https://github.com/alexgkendall/SegNet-Tutorial/tree/master/CamVid
python visualizeDataset.py
: Visual samples
python train.py
: Execution train
python predict.py
: Execution predict
You can modify the parameter in project switching model or cloning the historical version.
Visualization results:
Visualization results: