metric_utils.py [314:386]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
        return self.measure()

class PSNRMeter:

    def __init__(self, size=800):
        self.results = []
        self.size = size

    def read_img_list(self, img_list):
        size = self.size
        images = []
        white_background = np.ones((size, size, 3), dtype=np.uint8) * 255
        for img_path in img_list:
            img = cv2.imread(img_path, cv2.IMREAD_UNCHANGED)

            if img.shape[2] == 4:  # Handle BGRA images
                alpha = img[:, :, 3]  # Extract alpha channel
                img = cv2.cvtColor(img,
                                   cv2.COLOR_BGRA2BGR)  # Convert BGRA to BGR

                img = cv2.cvtColor(img,
                                   cv2.COLOR_BGR2RGB)  # Convert BGR to RGB
                img[np.where(alpha == 0)] = [
                    255, 255, 255
                ]  # Set transparent pixels to white
            else:  # Handle other image formats like JPG and PNG
                img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)

            img = cv2.resize(img, (size, size), interpolation=cv2.INTER_AREA)
            images.append(img)

        images = np.stack(images, axis=0)
        # images[np.where(images == 0)] = 255  # Set black pixels to white
        # images = np.where(images == 0, white_background, images)  # Set transparent pixels to white
        images = images.astype(np.float32) / 255.0
        # print(images.shape)
        return images

    def update(self, preds, truths):
        # print(preds.shape)

        psnr_values = []
        # For each pair of images in the batches
        for img1, img2 in zip(preds, truths):
            # Compute the PSNR and add it to the list
            # print(img1.shape,img2.shape)

            # for debug
            # plt.imshow(img1)
            # plt.show()
            # plt.imshow(img2)
            # plt.show()

            psnr = compare_psnr(
                img1, img2,
                data_range=1.0)  # assuming your images are scaled to [0,1]
            # print(f"temp psnr {psnr}")
            psnr_values.append(psnr)

        # Convert the list of PSNR values to a numpy array
        self.results = psnr_values

    def measure(self):
        return np.mean(self.results)

    def report(self):
        return f'PSNR = {self.measure():.6f}'

    # * recommend to use this function for evaluation
    def score_gt(self, ref_paths, novel_paths):
        self.results = []
        # [B, N, 3] or [B, H, W, 3], range[0, 1]
        preds = self.read_img_list(ref_paths)
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



threestudio/scripts/metric_utils.py [238:311]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
        return self.measure()


class PSNRMeter:

    def __init__(self, size=800):
        self.results = []
        self.size = size

    def read_img_list(self, img_list):
        size = self.size
        images = []
        white_background = np.ones((size, size, 3), dtype=np.uint8) * 255
        for img_path in img_list:
            img = cv2.imread(img_path, cv2.IMREAD_UNCHANGED)

            if img.shape[2] == 4:  # Handle BGRA images
                alpha = img[:, :, 3]  # Extract alpha channel
                img = cv2.cvtColor(img,
                                   cv2.COLOR_BGRA2BGR)  # Convert BGRA to BGR

                img = cv2.cvtColor(img,
                                   cv2.COLOR_BGR2RGB)  # Convert BGR to RGB
                img[np.where(alpha == 0)] = [
                    255, 255, 255
                ]  # Set transparent pixels to white
            else:  # Handle other image formats like JPG and PNG
                img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)

            img = cv2.resize(img, (size, size), interpolation=cv2.INTER_AREA)
            images.append(img)

        images = np.stack(images, axis=0)
        # images[np.where(images == 0)] = 255  # Set black pixels to white
        # images = np.where(images == 0, white_background, images)  # Set transparent pixels to white
        images = images.astype(np.float32) / 255.0
        # print(images.shape)
        return images

    def update(self, preds, truths):
        # print(preds.shape)

        psnr_values = []
        # For each pair of images in the batches
        for img1, img2 in zip(preds, truths):
            # Compute the PSNR and add it to the list
            # print(img1.shape,img2.shape)

            # for debug
            # plt.imshow(img1)
            # plt.show()
            # plt.imshow(img2)
            # plt.show()

            psnr = compare_psnr(
                img1, img2,
                data_range=1.0)  # assuming your images are scaled to [0,1]
            # print(f"temp psnr {psnr}")
            psnr_values.append(psnr)

        # Convert the list of PSNR values to a numpy array
        self.results = psnr_values

    def measure(self):
        return np.mean(self.results)

    def report(self):
        return f'PSNR = {self.measure():.6f}'

    # * recommend to use this function for evaluation
    def score_gt(self, ref_paths, novel_paths):
        self.results = []
        # [B, N, 3] or [B, H, W, 3], range[0, 1]
        preds = self.read_img_list(ref_paths)
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -