Function val_single_volume

utils/utils.py:239–274

(image, label, net, classes, patch_size=[256, 256], test_save_path=None, case=None, z_spacing=1)

    return metric_list

def val_single_volume(image, label, net, classes, patch_size=[256, 256], test_save_path=None, case=None, z_spacing=1):
    image, label = image.squeeze(0).cpu().detach().numpy(), label.squeeze(0).cpu().detach().numpy()

    if len(image.shape) == 3:
        prediction = np.zeros_like(label)
        for ind in range(image.shape[0]):
            slice = image[ind, :, :]
            x, y = slice.shape[0], slice.shape[1]
            if x != patch_size[0] or y != patch_size[1]:
                slice = zoom(slice, (patch_size[0] / x, patch_size[1] / y), order=3)  # previous using 0
            input = torch.from_numpy(slice).unsqueeze(0).unsqueeze(0).float().cuda()
            net.eval()
            with torch.no_grad():
                P = net(input)
                outputs = 0.0
                outputs = P[-1]
                out = torch.argmax(torch.softmax(outputs, dim=1), dim=1).squeeze(0)
                out = out.cpu().detach().numpy()
                if x != patch_size[0] or y != patch_size[1]:
                    pred = zoom(out, (x / patch_size[0], y / patch_size[1]), order=0)
                else:
                    pred = out
                prediction[ind] = pred
    else:
        input = torch.from_numpy(image).unsqueeze(
            0).unsqueeze(0).float().cuda()
        net.eval()
        with torch.no_grad():
            P = net(input)
            outputs = P[-1]
            out = torch.argmax(torch.softmax(outputs, dim=1), dim=1).squeeze(0)
            prediction = out.cpu().detach().numpy()
    metric_list = []
    for i in range(1, classes):
        metric_list.append(calculate_dice_percase(prediction == i, label == i))
    return metric_list

def horizontal_flip(image):
    image = image[:, ::-1, :]