INNER CODE UNIT · Python

labels

clovaai/deep-text-recognition-benchmark · dataset.py:299

        images, labels = zip(*batch)

        if self.keep_ratio_with_pad:  # same concept with 'Rosetta' paper
            resized_max_w = self.imgW
            input_channel = 3 if images[0].mode == 'RGB' else 1
            transform = NormalizePAD((input_channel, self.imgH, resized_max_w))

            resized_images = []
            for image in images:
                w, h = image.size
                ratio = w / float(h)
                if math.ceil(self.imgH * ratio) > self.imgW:
                    resized_w = self.imgW
                else:
                    resized_w = math.ceil(self.imgH * ratio)

                resized_image = image.resize((resized_w, self.imgH), Image.BICUBIC)
                resized_images.append(transform(resized_image))

View source record →