Skip to content

Kathange/skeleton_detection_similarity

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

51 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

skeleton_detection_similarity

This is a series of skeletal similarity detections between two videos.

The only two program that can execute the skeleton_detection_similarity is j.py & k.py.

Package

pip install mediapipe numpy opencv-python

Code Explain

Program Explain
a.py Test whether the mediapipe can execute normally
b.py How to extract the skeleton for angle calculations
Determine current posture by calculating angles
c.py Use OpenCV to read video
d.py Use OpenCV to read two video
e.py Use OpenCV to read two video
One on the left and one on the right
f.py Use OpenCV to read video
Do skeleton detect
g.py Use OpenCV to read two video
One on the left and one on the right
Do skeleton detect
Currently unavailable
Because mediapipe does not support detecting the skeletons of multiple people
h.py Use OpenCV to read two video
Do skeleton detect separately
There is two videos so we have two windows
i.py Use OpenCV to read two video
Do skeleton detect separately
After analysis, merge into the same window for viewing
j.py Use OpenCV to read two video
Do skeleton detect separately
And analyze the skeleton similarity
After analysis, merge into the same window for viewing
k.py Use OpenCV to read two video
Do skeleton detect separately
And analyze the skeleton similarity
After analysis, merge into the same window for viewing
Turn j.py into a class

Demo

Take j.py as an example :

demo

After the pre-processing is set, a video with skeleton analysis will pop up.

After detection, you can also see the skeleton analysis similarity percentage on the terminal. The similarity is 95.24%.

References

使用Mediapipe分析動作 reference of a.py & b.py
Mediapipe姿勢偵測Pose
B3.mp4 & B4.mp4 from YT 5:44
B5.mp4 & B6.mp4 from YT 2:40
B7.mp4 & B8.mp4 from YT 5:40

Releases

No releases published

Packages

No packages published

Languages