UHow commited on
Commit
82285bb
·
verified ·
1 Parent(s): 24fca51

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +1 -13
README.md CHANGED
@@ -24,7 +24,7 @@ datasets:
24
  - `2025/08/15`: We released our annotated DanceTrack-L, MOT17-L, and SportsMOT-L.
25
  - `2025/05/22`: Our work has been accepted by IEEE TCSVT.
26
  - `2024/06/14`: We released our trained models.
27
- - `2024/06/12`: We released our code.
28
  - `2024/06/11`: We released our technical report on [arxiv](https://arxiv.org/abs/2406.04844.pdf). Our code and models are coming soon!
29
 
30
  <br>
@@ -53,21 +53,9 @@ at both scene and instance levels.
53
 
54
  - **LG-MOT** is a new multi-object tracking framework which leverages language information at different granularity during training to enhance object association capabilities. During training, our **ISG** module aligns each node embedding $\phi(b_i^k)$ with instance-level descriptions embeddings $\varphi_i$, while our **SPG** module aligns edge embeddings $\hat{E}_{(u,v)}$ with scene-level descriptions embeddings $\varphi_s$ to guide correlation estimation after message passing. Our approach does not require language description during inference.
55
 
56
- <div align=center>
57
- <img src="source/LG-MOT_overview.png" width="90%"/>
58
- </div>
59
 
60
  - **LG-MOT** increases 1.2% in terms of IDF1 over the baseline SUSHI intra-domain, while significantly improves 11.2% in cross-domain.
61
 
62
- <div align=center>
63
- <img src="source/performance.png" width="50%"/>
64
- </div>
65
- <!-- <img src="source/performance.png" width="50%"/> -->
66
-
67
- ## Visualization
68
- <div align=center>
69
- <img src="source/visualization.png" width="90%"/>
70
- </div>
71
 
72
  ## Performance on Benchmarks
73
  ### MOT17 Challenge - Test Set
 
24
  - `2025/08/15`: We released our annotated DanceTrack-L, MOT17-L, and SportsMOT-L.
25
  - `2025/05/22`: Our work has been accepted by IEEE TCSVT.
26
  - `2024/06/14`: We released our trained models.
27
+ - `2024/06/12`: We released our code on [github](https://github.com/WesLee88524/LG-MOT).
28
  - `2024/06/11`: We released our technical report on [arxiv](https://arxiv.org/abs/2406.04844.pdf). Our code and models are coming soon!
29
 
30
  <br>
 
53
 
54
  - **LG-MOT** is a new multi-object tracking framework which leverages language information at different granularity during training to enhance object association capabilities. During training, our **ISG** module aligns each node embedding $\phi(b_i^k)$ with instance-level descriptions embeddings $\varphi_i$, while our **SPG** module aligns edge embeddings $\hat{E}_{(u,v)}$ with scene-level descriptions embeddings $\varphi_s$ to guide correlation estimation after message passing. Our approach does not require language description during inference.
55
 
 
 
 
56
 
57
  - **LG-MOT** increases 1.2% in terms of IDF1 over the baseline SUSHI intra-domain, while significantly improves 11.2% in cross-domain.
58
 
 
 
 
 
 
 
 
 
 
59
 
60
  ## Performance on Benchmarks
61
  ### MOT17 Challenge - Test Set