# FCN-pytorch

**Repository Path**: dalaska/FCN-pytorch

## Basic Information

- **Project Name**: FCN-pytorch
- **Description**:  Easiest Fully Convolutional Networks
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 1
- **Created**: 2019-12-06
- **Last Updated**: 2020-12-18

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

[![Open Source Love](https://badges.frapsoft.com/os/v1/open-source-150x25.png?v=103)](https://github.com/ellerbrock/open-source-badges/)

## 🚘 The easiest implementation of fully convolutional networks

- Task: __semantic segmentation__, it's a very important task for automated driving

- The model is based on CVPR '15 best paper honorable mentioned [Fully Convolutional Networks for Semantic Segmentation](https://arxiv.org/abs/1411.4038)

## Results
### Trials
<img align='center' style="border-color:gray;border-width:2px;border-style:dashed"   src='result/trials.png' padding='5px' height="150px"></img>

### Training Procedures
<img align='center' style="border-color:gray;border-width:2px;border-style:dashed"   src='result/result.gif' padding='5px' height="150px"></img>


## Performance

I train with two popular benchmark dataset: CamVid and Cityscapes

|dataset|n_class|pixel accuracy|
|---|---|---
|Cityscapes|20|96%
|CamVid|32|93%

## Training

### Install packages
```bash
pip3 install -r requirements.txt
```

and download pytorch 0.2.0 from [pytorch.org](pytorch.org)

and download [CamVid](http://mi.eng.cam.ac.uk/research/projects/VideoRec/CamVid/) dataset (recommended) or [Cityscapes](https://www.cityscapes-dataset.com/) dataset

### Run the code
- default dataset is CamVid

create a directory named "CamVid", and put data into it, then run python codes:
```python
python3 python/CamVid_utils.py 
python3 python/train.py CamVid
```

- or train with CityScapes

create a directory named "CityScapes", and put data into it, then run python codes:
```python
python3 python/CityScapes_utils.py 
python3 python/train.py CityScapes
```

## Author
Po-Chih Huang / [@pochih](https://pochih.github.io/)