evaluate.py

import torch
import torch.nn.functional as F
from tqdm import tqdm
from torchvision import transforms

from utils.dice_score import multiclass_dice_coeff, dice_coeff


def evaluate(net, dataloader, device, pretrained, test=False):
    net.eval()
    num_val_batches = len(dataloader)
    dice_score = 0

    # iterate over the validation set
    for batch in tqdm(dataloader, total=num_val_batches, desc='Validation round', unit='batch', leave=False):
        image, mask_true = batch['image'], batch['mask']
        # move images and labels to correct device and type
        image = image.to(device=device, dtype=torch.float32)
        mask_true = mask_true.to(device=device, dtype=torch.long)
        if pretrained:
            mask_true = F.one_hot(mask_true, 2).permute(0, 3, 1, 2).float()
        else:
            mask_true = F.one_hot(mask_true, net.n_classes).permute(0, 3, 1, 2).float()

        with torch.no_grad():
            # predict the mask
            mask_pred = net(image)

            if test:
                transform = transforms.ToPILImage()
                print(mask_pred[0])
                temp = transform(mask_pred[0]).convert('L')
                temp.save('.\data\PREDICT\preds/' + str(batch) + '.tif')
            # convert to one-hot format
            # if net.n_classes == 1:
            #     mask_pred = (F.sigmoid(mask_pred) > 0.5).float()
            #     # compute the Dice score
            #     dice_score += dice_coeff(mask_pred, mask_true, reduce_batch_first=False)
            if pretrained:
                mask_pred = F.one_hot(mask_pred.argmax(dim=1), 2).permute(0, 3, 1, 2).float()
            else:
                mask_pred = F.one_hot(mask_pred.argmax(dim=1), net.n_classes).permute(0, 3, 1, 2).float()
            # compute the Dice score, ignoring background
            dice_score += multiclass_dice_coeff(mask_pred[:, 1:, ...], mask_true[:, 1:, ...], reduce_batch_first=False)

           

    net.train()

    # Fixes a potential division by zero error
    if num_val_batches == 0:
        return dice_score
    return dice_score / num_val_batches