eval_Kvasir.py

import os
import torch
import torch.nn as nn
import torch.nn.functional as F
import torchvision.transforms as transforms
from PIL import Image
import numpy as np
import pandas
import cv2

class test_dataset:
    def __init__(self, image_root, gt_root):
        self.images = [image_root + f for f in os.listdir(image_root) if f.endswith('.jpg') or f.endswith('.png')]
        self.gts = [gt_root + f for f in os.listdir(gt_root) if f.endswith('.jpg') or f.endswith('.png')]
        self.images = sorted(self.images)
        self.gts = sorted(self.gts)
        self.transform = transforms.ToTensor()
        self.gt_transform = transforms.ToTensor()
        self.size = len(self.images)
        self.index = 0

    def load_data(self):
        image = self.rgb_loader(self.images[self.index])
        image = self.transform(image).unsqueeze(0)
        gt = self.binary_loader(self.gts[self.index])
        name = self.images[self.index].split('/')[-1]
        if name.endswith('.jpg'):
            name = name.split('.jpg')[0] + '.png'
        self.index += 1
        return image, gt, name

    def rgb_loader(self, path):
        with open(path, 'rb') as f:
            img = Image.open(f)
            return img.convert('RGB')

    def binary_loader(self, path):
        with open(path, 'rb') as f:
            img = Image.open(f)
            return img.convert('L')


if __name__ == '__main__':
    for _data_name in ['Kvasir']:
        ###########################################################
        ##### image_root : your model inference results' path   ###
        ##### gt_root : gt files' path                          ###
        ###########################################################
        image_root = '/home/james128333/HarDNet-MSEG/results/HarDMSEG/Kvasir/'
        gt_root = '/work/james128333/PraNet/TestDataset/Kvasir/masks/'
    
        test_loader = test_dataset(image_root, gt_root)
        b=0.0
    for i in range(test_loader.size):
        image, gt, name = test_loader.load_data()
        gt = np.asarray(gt, np.float32)
        gt /= (gt.max() + 1e-8)
        image = image
        input = image[0,1,:,:]
        input = np.array(input)
        
        target = np.array(gt)
        N = gt.shape
        smooth = 1
        
        
        input_flat = np.reshape(input,(-1))
        target_flat = np.reshape(target,(-1))
        
        intersection = (input_flat*target_flat)
        #intersection = (iflat * tflat).sum()
        #A_sum = input_flat.sum()
        #B_sum = target_flat.sum()
        #intersection = (input_flat * target_flat).sum()
        #a=  ((2. * intersection + smooth) / (A_sum + B_sum + smooth) )
        #loss = 2 * (intersection.sum(1) + smooth) / (input_flat.sum(1) + target_flat.sum(1) + smooth)
        loss =  (2 * intersection.sum() + smooth) / (input_flat.sum() + target_flat.sum() + smooth)
        #loss =  loss.sum() / N[1]
        a =  '{:.4f}'.format(loss)
        a = float(a)
        
        b = b + a
        print( i, a)
    print(b/test_loader.size)