I want to learn with YOLOv3.
At that time, the class number and coordinates are given, but can I add the distance to the camera and the size of the object here? I think.
If so, which file in the src folder should I play with?

  • Answer # 1


    Can you add the distance to the camera and the size of the object?

    I think that depth information is necessary for the distance from the camera and the size of the object, such as with a twin-lens camera.

    Given the depth information, it is the draw_detection function of image.c that draws the rectangle determined by YOLO, so I think that it is better to calculate the depth and size at that side and draw it.

  • Answer # 2

    Thank you for your answer.
    I'm wondering if we can make a system that gives us the distance between the camera and the object using a web camera that can be purchased easily.
    So, is it possible to estimate the distance from the size of the bounding box to be created if the distance of the object is learned from the beginning? Is it impossible?