updated readme with ntu dataset and license detail

README.md

@@ -102,6 +102,7 @@ The MVTamperBench dataset integrates videos from several sources, each contribut
 | MoVQA     | Diverse scenes for scene transition comprehension                   |
 | VLN-CE       | Indoor navigation from agent perspective                            |
 | TVQA     | TV show scenes for episodic reasoning                               |
+| NTU RGB+D | Used for human action recognition                                  |
 
 ### Dataset Expansion
 
@@ -112,7 +113,7 @@ The original MVBench dataset contains 3,699 videos, which have been systematical
 
 Below is a visual representation of the tampered video length distribution:
 
-![Tampered Video Length Distribution](./assert/tampered_video_length_distribution.png "Distribution of tampered video lengths")
+![Tampered Video Length Distribution](path/to/tampered_video_length_distribution.png "Distribution of tampered video lengths")
 
 ---
 
@@ -180,4 +181,4 @@ If you use MVTamperBench in your research, please cite:
 
 ## License
 
-MVTamperBench is released under the MIT License. See `LICENSE` for details.
+MVTamperBench is built upon MVBench and therefore operates under the same license as the original MVBench. For more details, please refer to the MVBench README.