Image Editing only works for fixed size inputs?

Hello, I am trying to use DeepAR’s webkit SDK on singular images instead of the web camera, but can only get images of fixed-size to work. Is this expected?

I am following this repository:

If I even slightly change the image size, no face is detected. How might I go about fixing this?


it wll work best with the same dimensions, although it should still work with different size images, but you may need to enqueue the same image a few times using processImage for the tracking to catch up.