File size: 2,582 Bytes
d872151
 
077a7c2
d872151
077a7c2
 
 
 
 
 
 
 
 
 
 
 
9adbd65
 
7f58721
d872151
 
 
 
 
 
4261796
d872151
 
d9d6736
d872151
077a7c2
29d5f12
d872151
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ac3ddc8
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
# LISA: Reasoning Segmentation Via Large Language Model

This is the official implementation of ***LISA (large Language Instructed Segmentation Assistant)***. 

## News
- [x] [2023.8.2] Paper is released and github repo is created.

## TODO 
- [ ] Huggingface Demo
- [ ] ReasonSeg Dataset Release
- [ ] Codes and models Release

LISA can handle cases involving: 1) complex reasoning; 2) world knowledge; 3) explanatory answers; 4) multi-turn conversation.It demonstrates robust zero-shot capability when trained exclusively on reasoning-free datasets.
<p align="center"> <img src="imgs/fig_teaser4_crop.png" width="100%"> </p>

## Abstract
In this work, we propose a new segmentation task --- ***reasoning segmentation***. The task is designed to output a segmentation mask given a complex and implicit query text. We establish a benchmark comprising over one thousand image-instruction pairs, incorporating intricate reasoning and world knowledge for evaluation purposes. Finally, we present LISA: Large-language Instructed Segmentation Assistant, which inherits the language generation capabilities of the multi-modal Large Language Model (LLM) while also possessing the ability to produce segmentation masks.
For more details, please refer to:

**LISA: Reasoning Segmentation Via Large Language Model [[Paper]()]** <br />
[Xin Lai](https://scholar.google.com/citations?user=tqNDPA4AAAAJ&hl=zh-CN),
[Zhuotao Tian](https://scholar.google.com/citations?user=mEjhz-IAAAAJ&hl=en),
[Yukang Chen](https://scholar.google.com/citations?user=6p0ygKUAAAAJ&hl=en),
[Yanwei Li](https://scholar.google.com/citations?user=I-UCPPcAAAAJ&hl=zh-CN),
[Yuhui Yuan](https://scholar.google.com/citations?user=PzyvzksAAAAJ&hl=en),
[Shu Liu](https://scholar.google.com.hk/citations?user=BUEDUFkAAAAJ&hl=zh-CN),
[Jiaya Jia](https://scholar.google.com/citations?user=XPAkzTEAAAAJ&hl=en)<br />

<p align="center"> <img src="imgs/fig_overview_v6_crop.png" width="100%"> </p>

## Experimental results
<p align="center"> <img src="imgs/Table1.png" width="80%"> </p>

## Citation 
If you find this project useful in your research, please consider citing:

```
@article{reason-seg,
  title={LISA: Reasoning Segmentation Via Large Language Model},
  author={Xin Lai and Zhuotao Tian and Yukang Chen and Yanwei Li and Yuhui Yuan and Shu Liu and Jiaya Jia},
  journal={arXiv:},
  year={2023}
}

```

## Acknowledgement
-  This work is built upon the [LLaMA](https://github.com/facebookresearch/llama), [SAM](https://github.com/facebookresearch/segment-anything), and [LLaVA](https://github.com/haotian-liu/LLaVA).