Clarification on the sample submission/Docker script available @ProstateCancerRiskPrediction

Hi, This is my first time working with docker files, I had a doubt regarding the process.py script available here: https://github.com/chaimeleon-eu/OpenChallenge/tree/master/ProstateCancerRiskPrediction In the Prostatecancerriskprediction class definition the following line saves only the first image in the input_dir # path to image file self.image_input_dir = "/input/images/axial-t2-prostate-mri/" self.image_input_path = list(Path(self.image_input_dir).glob("*.mha"))**[0]** Therefore I am not sure how to handle multiple input files if they are present in the input_dir . Will the image data be replaced at multiple calls of the Prostatecancerriskprediction().predict() ? Or should I remove the [0] index and proceed with loop to iterate over every .mha file?

Best, Dhruv

Hi, the inference is case by case, that is, you have to prepare a docker that does inference on an image only and generate the inference on that case. The grand-challenge platform is the one in charge of applying the algorithm that you upload on all data.

Hope we could help.

Hi, Thank you! That clears my doubt