Abstract

This publication presents object detection between objects and the hand of a person using mediapipe to get the hand gestures and yolo to detect the object

Methodology

the project works in a flow

initial processing:

    getting the image from a real-time stream or a recorded video 
    detecting the hand and from that getting the hand points once 
    they are in place yolo detects objects around the objects that are 
     to  detect can be controlled by the user.

detecting distance:

   detecting the distance is the hard part hence both the Yolo and 
   media pipe does not provide any useable info for this so we just
   mark a pixel point on both hands and the object after that we 
   calculate the distance between them pixel-based. 

Results

the algorithm was successful in detecting the distance. 

Vedio on Linkedin