How to extract object orientation from video stream (Webcam) - Python

Question

I found this guide which teaches how to refine the orientation of objects from images. I would love to know if it can and should be used to analyze the orientation of objects displayed in video streams.

The basis for the work is from the scientific publication found in this video . I want to know how they got information about the direction of the Fish's face.

Thanks , Avishai

Answer 1

You will probably need library like opencv to get orientation information from the image. You can apply threshold after converting this image to grayscale and extract contour of the image. After that you need to follow something like below pattern to get orientation. Very easy, just a little bit search you can find a lot of similar examples as well.

rectangle_for_angle = cv2.minAreaRect(cntrs[0])
angle = rectangle_for_angle[-1]
rect_points = cv2.boxPoints(rectangle_for_angle)
rect_points_result = np.int0(rect_points)
#You can also draw rotated image
cv2.drawContours(image,[rect_points_result],0,(0,0,255),2)