Is there some established Object Detection algorithm that is able to detect the four corners of an arbitrary quadrilateral (x0,y0,x1,y1,x2,y2,x3,y3) as opposed to the more typical perpendicular rectangular (x,y,w,h) ?


You can use OpenCV's cv2.minAreaRect() to detect oriented/rotated rectangular bounding boxes. Below's an example result from OpenCV-Python-tutorials:

Using cv2.minAreaRect()

Alternatively, you could train a supervised object detection model to output 8 co-ordinate values (x0,y0,x1,y1,x2,y2,x3,y3) of the quadrilateral by training with a labeled oriented-bounding-box dataset. You could also create the bounding box labels yourself for the same by using tools such as VGG Annotator Tool among others.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.