Intersection over Union for Python [ Keras ]

computeIoU.py

import numpy as np

def computeIoU(y_pred_batch, y_true_batch):
    return np.mean(np.asarray([pixelAccuracy(y_pred_batch[i], y_true_batch[i]) for i in range(len(y_true_batch))])) 

def pixelAccuracy(y_pred, y_true):
    y_pred = np.argmax(np.reshape(y_pred,[N_CLASSES_PASCAL,img_rows,img_cols]),axis=0)
    y_true = np.argmax(np.reshape(y_true,[N_CLASSES_PASCAL,img_rows,img_cols]),axis=0)
    y_pred = y_pred * (y_true>0)
    
    return 1.0 * np.sum((y_pred==y_true)*(y_true>0)) /  np.sum(y_true>0)

How does this actually work in the scope of a full model? Also this does not look like IOU. Can u explain more??

i want to know what is the different between mean IU and dice?