consult about the input data

Dear author:
I didn't find any sparse perspective images in the source code. I only found some json data, as well as condition and mask videos. Could you please tell me where to set the input data?

<img width="818" height="1342" alt="Image" src="https://github.com/user-attachments/assets/5a988033-80cd-4416-a43f-21edc5fa68b1" />