Home

Awesome

FastInst: A Simple Query-Based Model for Real-Time Instance Segmentation

[arXiv] [BibTeX]

<p align="center"><img width="100%" src="figures/fastinst.png" /></p>

Features


Updates

Installation

See installation instructions.

Getting Started

See Results.

See Preparing Datasets for FastInst.

See Getting Started.


Results

<p><img width="50%" src="figures/trade-off.png" /></p>

COCO Instance Segmentation

<table><tbody> <!-- START TABLE --> <!-- TABLE HEADER --> <th valign="bottom">Name</th> <th valign="bottom">Backbone</th> <th valign="bottom">Epochs</th> <th valign="bottom">Input</th> <th valign="bottom">AP<sup>val</sup></th> <th valign="bottom">AP</th> <th valign="bottom">Params</th> <th valign="bottom">GFlops</th> <th valign="bottom">FPS (V100)</th> <th valign="bottom">download</th> <tr> <td align="left"><a href="configs/coco/instance-segmentation/fastinst_R50_ppm-fpn_x1_576.yaml">FastInst-D1</a></td> <td align="center">R50</td> <td align="center">50</td> <td align="center">576</td> <td align="center">34.9</td> <td align="center">35.6</td> <td align="center">30M</td> <td align="center">49.6</td> <td align="center">53.8</td> <td align="center"><a href="https://github.com/junjiehe96/FastInst/releases/download/v0.1.0/fastinst_R50_ppm-fpn_x1_576_34.9.pth">model</a></td> </tr> <tr> <td align="left"><a href="configs/coco/instance-segmentation/fastinst_R50_ppm-fpn_x3_640.yaml">FastInst-D3</a></td> <td align="center">R50</td> <td align="center">50</td> <td align="center">640</td> <td align="center">37.9</td> <td align="center">38.6</td> <td align="center">34M</td> <td align="center">75.5</td> <td align="center">35.5</td> <td align="center"><a href="https://github.com/junjiehe96/FastInst/releases/download/v0.1.0/fastinst_R50_ppm-fpn_x3_640_37.9.pth">model</a></td> </tr> <tr> <td align="left"><a href="configs/coco/instance-segmentation/fastinst_R101_ppm-fpn_x3_640.yaml">FastInst-D3</a></td> <td align="center"><a href="https://dl.fbaipublicfiles.com/detectron2/ImageNetPretrained/MSRA/R-101.pkl">R101</a></td> <td align="center">50</td> <td align="center">640</td> <td align="center">38.9</td> <td align="center">39.9</td> <td align="center">53M</td> <td align="center">112.9</td> <td align="center">28.0</td> <td align="center"><a href="https://github.com/junjiehe96/FastInst/releases/download/v0.1.0/fastinst_R101_ppm-fpn_x3_640_38.9.pth">model</a></td> </tr> <tr> <td align="left"><a href="configs/coco/instance-segmentation/fastinst_R50-vd-dcn_ppm-fpn_x1_576.yaml">FastInst-D1</a></td> <td align="center"><a href="https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/resnet50d_ra2-464e36ba.pth">R50-d-DCN</a></td> <td align="center">50</td> <td align="center">576</td> <td align="center">37.4</td> <td align="center">38.0</td> <td align="center">30M</td> <td align="center"> - </td> <td align="center">47.8</td> <td align="center"><a href="https://github.com/junjiehe96/FastInst/releases/download/v0.1.0/fastinst_R50-vd-dcn_ppm-fpn_x1_576_37.4.pth">model</a></td> </tr> <tr> <td align="left"><a href="configs/coco/instance-segmentation/fastinst_R50-vd-dcn_ppm-fpn_x3_640.yaml">FastInst-D3</a></td> <td align="center"><a href="https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/resnet50d_ra2-464e36ba.pth">R50-d-DCN</a></td> <td align="center">50</td> <td align="center">640</td> <td align="center">40.1</td> <td align="center">40.5</td> <td align="center">35M</td> <td align="center"> - </td> <td align="center">32.5</td> <td align="center"><a href="https://github.com/junjiehe96/FastInst/releases/download/v0.1.0/fastinst_R50-vd-dcn_ppm-fpn_x3_640_40.1.pth">model</a></td> </tr> </tbody></table>

Getting Started

This document provides a brief intro of the usage of FastInst.

Please see Getting Started with Detectron2 for full usage.

Evaluate our pretrained models

Train FastInst to reproduce results

LICNESE

FastInst is released under the MIT Licence.

<a name="CitingFastInst"></a>Citing FastInst

If you find FastInst is useful in your research or applications, please consider giving us a star 🌟 and citing FastInst by the following BibTeX entry.

@article{he2023fastinst,
  title={FastInst: A Simple Query-Based Model for Real-Time Instance Segmentation},
  author={He, Junjie and Li, Pengyu and Geng, Yifeng and Xie, Xuansong},
  journal={arXiv preprint arXiv:2303.08594},
  year={2023}
}

Acknowledgement

Sincerely thanks to these excellent opensource projects