Exemplo n.º 1
0
    def get_data(self, img_path):
        """
        1. load image from img_path.
        2. resize or oversampling.
        3. transformer data: transpose, channel swap, sub mean.
        return K x H x W ndarray.

        img_path: image path.
        """
        image = image_util.load_image(img_path, self.is_color)
        # Another way to extract oversampled features is that
        # cropping and averaging from large feature map which is
        # calculated by large size of image.
        # This way reduces the computation.
        if self.oversample:
            # image_util.resize_image: short side is self.resize_dim
            image = image_util.resize_image(image, self.resize_dim)
            image = np.array(image)
            input = np.zeros((1, image.shape[0], image.shape[1], 3),
                             dtype=np.float32)
            input[0] = image.astype(np.float32)
            input = image_util.oversample(input, self.crop_dims)
        else:
            image = image.resize(self.crop_dims, Image.ANTIALIAS)
            input = np.zeros((1, self.crop_dims[0], self.crop_dims[1], 3),
                             dtype=np.float32)
            input[0] = np.array(image).astype(np.float32)

        data_in = []
        for img in input:
            img = self.transformer.transformer(img).flatten()
            data_in.append([img.tolist()])
        # paddle input: [[[]],[[]],...], [[]] is one sample.
        return data_in
Exemplo n.º 2
0
    def get_data(self, img_path):
        """
        1. load image from img_path.
        2. resize or oversampling.
        3. transformer data: transpose, channel swap, sub mean.
        return K x H x W ndarray.

        img_path: image path.
        """
        image = image_util.load_image(img_path, self.is_color)
        # Another way to extract oversampled features is that
        # cropping and averaging from large feature map which is
        # calculated by large size of image.
        # This way reduces the computation.
        if self.oversample:
            # image_util.resize_image: short side is self.resize_dim
            image = image_util.resize_image(image, self.resize_dim)
            image = np.array(image)
            input = np.zeros(
                (1, image.shape[0], image.shape[1], 3), dtype=np.float32)
            input[0] = image.astype(np.float32)
            input = image_util.oversample(input, self.crop_dims)
        else:
            image = image.resize(self.crop_dims, Image.ANTIALIAS)
            input = np.zeros(
                (1, self.crop_dims[0], self.crop_dims[1], 3), dtype=np.float32)
            input[0] = np.array(image).astype(np.float32)

        data_in = []
        for img in input:
            img = self.transformer.transformer(img).flatten()
            data_in.append([img.tolist()])
        # paddle input: [[[]],[[]],...], [[]] is one sample.
        return data_in
Exemplo n.º 3
0
    def get_data(self, img_path):
        """
        1. load image from img_path.
        2. resize or oversampling.
        3. transformer data: transpose, sub mean.
        return K x H x W ndarray.
        img_path: image path.
        """
        image = image_util.load_image(img_path, self.is_color)
        if self.oversample:
            # image_util.resize_image: short side is self.resize_dim
            image = image_util.resize_image(image, self.resize_dim)
            image = np.array(image)
            input = np.zeros((1, image.shape[0], image.shape[1], 3),
                             dtype=np.float32)
            input[0] = image.astype(np.float32)
            input = image_util.oversample(input, self.crop_dims)
        else:
            image = image.resize(self.crop_dims, Image.ANTIALIAS)
            input = np.zeros((1, self.crop_dims[0], self.crop_dims[1], 3),
                             dtype=np.float32)
            input[0] = np.array(image).astype(np.float32)

        data_in = []
        for img in input:
            img = self.transformer.transformer(img).flatten()
            data_in.append([img.tolist()])
        return data_in
Exemplo n.º 4
0
    def get_data(self, img_path):
        """
        1. load image from img_path.
        2. resize or oversampling.
        3. transformer data: transpose, sub mean.
        return K x H x W ndarray.
        img_path: image path.
        """
        image = image_util.load_image(img_path, self.is_color)
        if self.oversample:
            # image_util.resize_image: short side is self.resize_dim
            image = image_util.resize_image(image, self.resize_dim)
            image = np.array(image)
            input = np.zeros(
                (1, image.shape[0], image.shape[1], 3), dtype=np.float32)
            input[0] = image.astype(np.float32)
            input = image_util.oversample(input, self.crop_dims)
        else:
            image = image.resize(self.crop_dims, Image.ANTIALIAS)
            input = np.zeros(
                (1, self.crop_dims[0], self.crop_dims[1], 3), dtype=np.float32)
            input[0] = np.array(image).astype(np.float32)

        data_in = []
        for img in input:
            img = self.transformer.transformer(img).flatten()
            data_in.append([img.tolist()])
        return data_in