Inconsistences between the submitted results json and average scores.

Hi,

I submitted the predicted masks and received the result json. There are inconsistencies between the average result and the average obtained from the json of individual cases. Can you please ensure if the average is being calculated in the right way?

import json

with open('preddata.json') as jsonfile: data = json.load(jsonfile)

x_a = data.keys()

tumor = [] kidney = [] renalartery=[] renalvein = []

for idx, x in enumerate(xa): if idx > 0: # ignoring the average kidney.append(data[x]["Kidney"]["Dice"]) tumor.append(data[x]["Tumor"]["Dice"]) renalartery.append(data[x]["Renal Artery"]["Dice"]) renalvein.append(data[x]["Renal Vein"]["Dice"])

    print(x, "-", data[x]["Kidney"]["Dice"], "-", data[x]["Tumor"]["Dice"], 
          "-", data[x]["Renal Artery"]["Dice"], "-", data[x]["Renal Vein"]["Dice"])

import numpy as np print(np.mean(kidney)) print(np.mean(tumor)) print(np.mean(renalartery)) print(np.mean(renalvein))

Thanks!

Thanks for your message. We have fixed this bug and reuploaded the corrected evaluation code. Please resubmit your predicted masks.