Method __call__

yolox/data/data_augment.py:221–270

(self, image, targets, input_dim)

        self.max_labels = max_labels

    def __call__(self, image, targets, input_dim):
        boxes = targets[:, :4].copy()
        labels = targets[:, 4].copy()
        ids = targets[:, 5].copy()
        if len(boxes) == 0:
            targets = np.zeros((self.max_labels, 6), dtype=np.float32)
            image, r_o = preproc(image, input_dim, self.means, self.std)
            image = np.ascontiguousarray(image, dtype=np.float32)
            return image, targets

        image_o = image.copy()
        targets_o = targets.copy()
        height_o, width_o, _ = image_o.shape
        boxes_o = targets_o[:, :4]
        labels_o = targets_o[:, 4]
        ids_o = targets_o[:, 5]
        # bbox_o: [xyxy] to [c_x,c_y,w,h]
        boxes_o = xyxy2cxcywh(boxes_o)

        image_t = _distort(image)
        image_t, boxes = _mirror(image_t, boxes)
        height, width, _ = image_t.shape
        image_t, r_ = preproc(image_t, input_dim, self.means, self.std)
        # boxes [xyxy] 2 [cx,cy,w,h]
        boxes = xyxy2cxcywh(boxes)
        boxes *= r_

        mask_b = np.minimum(boxes[:, 2], boxes[:, 3]) > 1
        boxes_t = boxes[mask_b]
        labels_t = labels[mask_b]
        ids_t = ids[mask_b]

        if len(boxes_t) == 0:
            image_t, r_o = preproc(image_o, input_dim, self.means, self.std)
            boxes_o *= r_o
            boxes_t = boxes_o
            labels_t = labels_o
            ids_t = ids_o

        labels_t = np.expand_dims(labels_t, 1)
        ids_t = np.expand_dims(ids_t, 1)

        targets_t = np.hstack((labels_t, boxes_t, ids_t))
        padded_labels = np.zeros((self.max_labels, 6))
        padded_labels[range(len(targets_t))[: self.max_labels]] = targets_t[
            : self.max_labels
        ]
        padded_labels = np.ascontiguousarray(padded_labels, dtype=np.float32)
        image_t = np.ascontiguousarray(image_t, dtype=np.float32)
        return image_t, padded_labels


class ValTransform: