Update README.md

README.md

@@ -7,7 +7,7 @@ license: apache-2.0
 
 ### Dataset Description
 
-OSCBench is a benchmark dataset designed to evaluate **object state change (OSC)** reasoning in text-to-video (T2V) generation models. It provides structured prompts describing actions applied to objects (e.g., *peeling carrot*, *rolling dough*), where the correct outcome requires generating the appropriate **action-induced object state change**.
+OSCBench is a benchmark dataset designed to evaluate **object state change (OSC)** reasoning in text-to-video (T2V) generation models.
 
 OSCBench organizes prompts into three scenario types:
 
@@ -34,10 +34,15 @@ The OSCBench dataset contains **1,120 prompts** organized into three scenario ca
 - **Project Page:** https://hanxjing.github.io/OSCBench
 
 
-## Dataset Structure
+## Acknowledgements and Citation
 
-The dataset consists of **structured prompt files** describing actions applied to objects.
+If you find this dataset helpful, please consider citing the original work:
 
-Example prompt:
-
-A man is slicing apple in the kitchen.
+```bash
+@article{han2026oscbench,
+  title={OSCBench: Benchmarking Object State Change in Text-to-Video Generation},
+  author={Han, Xianjing and Zhu, Bin and Hu, Shiqi and Li, Franklin Mingzhe and Carrington, Patrick and Zimmermann, Roger and Chen, Jingjing},
+  journal={arXiv preprint arXiv:2603.11698},
+  year={2026}
+}
+```