https://github.com/aitorzip/keras-icnet

Keras implementation of Real-Time Semantic Segmentation on High-Resolution Images
https://github.com/aitorzip/keras-icnet

autonomous-driving computer-vision deep-learning fully-convolutional-networks image-processing image-segmentation keras semantic-segmentation tensorflow

Last synced: about 2 months ago
JSON representation

Keras implementation of Real-Time Semantic Segmentation on High-Resolution Images

• Host: GitHub
• URL: https://github.com/aitorzip/keras-icnet
• Owner: aitorzip
• License: mit
• Created: 2017-11-16T16:01:16.000Z (over 7 years ago)
• Default Branch: master
• Last Pushed: 2018-01-24T12:25:26.000Z (over 7 years ago)
• Last Synced: 2025-03-22T19:43:48.581Z (2 months ago)
• Topics: autonomous-driving, computer-vision, deep-learning, fully-convolutional-networks, image-processing, image-segmentation, keras, semantic-segmentation, tensorflow
• Language: Python
• Homepage: https://arxiv.org/abs/1704.08545
• Size: 25.2 MB
• Stars: 86
• Watchers: 8
• Forks: 33
• Open Issues: 11
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Keras-ICNet

### [[paper]](https://arxiv.org/abs/1704.08545)

Keras implementation of Real-Time Semantic Segmentation on High-Resolution Images. **Training in progress!**

    

## Requisites

- Python 3.6.3

- Keras 2.1.1 with Tensorflow backend

- A dataset, such as Cityscapes or Mapillary ([Mapillary](https://research.mapillary.com/) was used in this case).

## Train

Issue ```./train --help``` for options to start a training session, default arguments should work out-of-the-box.

You need to place the dataset following the next directory convention:

    .

    ├── mapillary                   

    |   ├── training

    |   |   ├── images             # Contains the input images

    |   |   └── instances          # Contains the target labels

    |   ├── validation

    |   |   ├── images

    |   |   └── instances

    |   └── testing

    |   |   └── images

    

These are the results of training for 300 epochs ```./train --epochs 300```

### Training

![conv6_cls_categorical_accuracy](https://raw.githubusercontent.com/ai-tor/Keras-ICNet/master/output/conv6_cls_categorical_accuracy.png)

![conv6_cls_loss](https://raw.githubusercontent.com/ai-tor/Keras-ICNet/master/output/conv6_cls_loss.png)

![loss](https://raw.githubusercontent.com/ai-tor/Keras-ICNet/master/output/loss.png)

### Validation

![val_conv6_cls_categorical_accuracy](https://raw.githubusercontent.com/ai-tor/Keras-ICNet/master/output/val_conv6_cls_categorical_accuracy.png)

![val_conv6_cls_loss](https://raw.githubusercontent.com/ai-tor/Keras-ICNet/master/output/val_conv6_cls_loss.png)

![val_loss](https://raw.githubusercontent.com/ai-tor/Keras-ICNet/master/output/val_loss.png)

## Test

Issue ```./test --help``` for options to start a testing session, default arguments should work out-of-the-box.

### Output examples

![10](https://raw.githubusercontent.com/ai-tor/Keras-ICNet/master/output/10.png)

![07](https://raw.githubusercontent.com/ai-tor/Keras-ICNet/master/output/7.png)

## TODO

* Perform class weighting