Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,34 @@
|
|
1 |
-
---
|
2 |
-
license: cc-by-nc-sa-4.0
|
3 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: cc-by-nc-sa-4.0
|
3 |
+
---
|
4 |
+
|
5 |
+
# SeC: Advancing Complex Video Object Segmentation via Progressive Concept Construction
|
6 |
+
|
7 |
+
[\[📂 GitHub\]](https://github.com/OpenIXCLab/SeC)
|
8 |
+
[\[📦 Model\]](https://huggingface.co/OpenIXCLab/SeC-4B)
|
9 |
+
[\[🌐 Homepage\]](https://rookiexiong7.github.io/projects/SeC/)
|
10 |
+
[\[📄 Paper\]](https://arxiv.org/abs/2507.xxxxx)
|
11 |
+
|
12 |
+
## Highlights
|
13 |
+
|
14 |
+
- 🔥We introduce **Segment Concept (SeC)**, a **concept-driven** segmentation framework for **video object segmentation** that integrates **Large Vision-Language Models (LVLMs)** for robust, object-centric representations.
|
15 |
+
- 🔥SeC dynamically balances **semantic reasoning** with **feature matching**, adaptively adjusting computational efforts based on **scene complexity** for optimal segmentation performance.
|
16 |
+
- 🔥We propose the **Semantic Complex Scenarios Video Object Segmentation (SeCVOS)** benchmark, designed to evaluate segmentation in challenging scenarios.
|
17 |
+
|
18 |
+
## SeCVOS Benchmark
|
19 |
+
|
20 |
+
We propose the Semantic Complex Scenarios Video Object Segmentation (SeCVOS) benchmark, specifically designed to assess a model’s ability to perform high-level semantic reasoning across complex visual narratives. SeCVOS contains 160 carefully curated multi-shot videos characterized by: 1) Highly discontinuous frame sequences, 2) Frequent reappearance of objects across disparate scenes, and 3) Abrupt shot transitions and dynamic camera motion.
|
21 |
+
|
22 |
+
| Benchmark | #Videos | Avg. Duration (s) | Disapp. Rate | Avg. #Scene |
|
23 |
+
| :------------------------------ | :------: | :---------------: | :----------: | :-----------: |
|
24 |
+
| DAVIS | 90 | 2.87 | 16.1% | 1.06 |
|
25 |
+
| YTVOS | 507 | 4.51 | 13.0% | 1.03 |
|
26 |
+
| MOSE | 311 | 8.68* | 41.5% | 1.06 |
|
27 |
+
| SA-V | 155 | 17.24 | 25.5% | 1.09 |
|
28 |
+
| LVOS | 140 | 78.36 | 7.8% | 1.47 |
|
29 |
+
| **SeCVOS (ours)** | 160 | 29.36 | 30.2% | **4.26** |
|
30 |
+
|
31 |
+
## License
|
32 |
+
Our annotations are licensed under a [CC BY-NC-SA 4.0 License](https://creativecommons.org/licenses/by-nc-sa/4.0/). They are available strictly for non-commercial research.
|
33 |
+
|
34 |
+
We uphold the rights of individuals and copyright holders. If you are featured in any of our video annotations or hold copyright to a video and wish to have its annotation removed from our dataset, please reach out to us. Send an email to [email protected] with the subject line beginning with SeCVOS, or raise an issue with the same title format. We commit to reviewing your request promptly and taking suitable action.
|