question for image size

[Sylvie-X]

I noticed that you resize input images for pre-processing in your user interface. Could you please confirm if you apply the same resizing technique when testing with DragBench?

[Sylvie-X]

I failed to obatin the same MD as those presented in your paper. I utilized the identical models, including Runway SD, and adhered to your code via the UI. The only difference lies in the resizing process. I resized both mask and image in accordance with your technique and also adjusted the points using the formula points = round(points * new_hw / old_hw).