Closed xuguozhi closed 4 years ago
您好,我在尝试用mnn C++做keypoint模型的c++移植,对于 这里的预处理不理解:https://github.com/610265158/Peppa_Pig_Face_Engine/blob/master/lib/core/api/face_landmark.py#L102 add = int(max(bbox_width, bbox_height)) bimg = cv2.copyMakeBorder(image, add, add, add, add, borderType=cv2.BORDER_CONSTANT, value=cfg.DATA.pixel_means) bbox += add
add = int(max(bbox_width, bbox_height))
bimg = cv2.copyMakeBorder(image, add, add, add, add, borderType=cv2.BORDER_CONSTANT, value=cfg.DATA.pixel_means)
bbox += add
请问能否协助指点C++的预处理吗?
已经解决
您好,我在尝试用mnn C++做keypoint模型的c++移植,对于 这里的预处理不理解:https://github.com/610265158/Peppa_Pig_Face_Engine/blob/master/lib/core/api/face_landmark.py#L102
add = int(max(bbox_width, bbox_height))
bimg = cv2.copyMakeBorder(image, add, add, add, add, borderType=cv2.BORDER_CONSTANT, value=cfg.DATA.pixel_means)
bbox += add
请问能否协助指点C++的预处理吗?