https://github.com/HKUSTCV/FSS-1000

FSS-1000, A 1000-class Dataset For Few-shot Segmentation
https://github.com/HKUSTCV/FSS-1000

Last synced: 3 months ago
JSON representation

FSS-1000, A 1000-class Dataset For Few-shot Segmentation

• Host: GitHub
• URL: https://github.com/HKUSTCV/FSS-1000
• Owner: HKUSTCV
• Created: 2019-04-06T10:19:13.000Z (over 5 years ago)
• Default Branch: master
• Last Pushed: 2020-12-16T07:03:09.000Z (almost 4 years ago)
• Last Synced: 2024-05-08T19:30:58.595Z (6 months ago)
• Language: Python
• Size: 11.2 MB
• Stars: 268
• Watchers: 9
• Forks: 47
• Open Issues: 13
• Metadata Files:
  • Readme: readme.md

Awesome Lists containing this project

README

        
# FSS-1000: A 1000 Class Dataset for Few-shot Segmentation 



We provide our dataset and PyTorch implementation for relation network benchmark. Details are in our [paper](https://arxiv.org/abs/1907.12347). 

## Prerequisites

- Linux or macOS

- Python 3

- CPU or NVIDIA GPU + CUDA CuDNN

- PyTorch 0.4+

## FSS-1000 Dataset

- Google drive: [download here](https://drive.google.com/open?id=16TgqOeI_0P41Eh3jWQlxlRXG9KIqtMgI)

- Online Preview: Coming soon

## Getting Started

### Testing

First, download pretrained model [here](https://drive.google.com/open?id=1Vk0Pq8vOZrfrDtCISMcJmAQnt9jkXfPn).

```

python autolabel.py -sd imgs/example/support -td imgs/example/query

```

- Set option ```-sd``` to the support directory and the script will input them as support set. 

- Set option ```-td``` to the path of your query images.

- Results will be saved under ```./results```

  

### Testing your own data

- Label 5 support images following the format in ```imgs/example/support/```.  

- Set your support and query path accordingly.

### Training

Arrange the dataset as described in ```get_oneshot_batch()``` in ```training.py```, then run

```

python training.py

```

## Citing

If you use this repository, dataset or want to reference our work, please use the following BibTeX entry.

```

@article{FSS1000,

Author = {Xiang Li and Tianhan Wei and Yau Pun Chen and Yu-Wing Tai and Chi-Keung Tang},

Title = {FSS-1000: A 1000-Class Dataset for Few-Shot Segmentation},

Year = {2020},

Journal = {CVPR},

}

```