Abstract:
Salat is the second pillar of Islam. In the Muslim community, it is the most essential
and fundamental worshiping activity that believers have to perform five times a day.
Salat consists ofseveral postures to perform with accuracy and needs to be performed
in a defined sequence. Besides spirituality, Salat has many benefits and welfares.
During learning many youngsters are hesitant to correct their postures or sometimes
they are unaware of mistakes, even elderly people have several physical constraints
and a very common issue ofremembering the sequence, count, and correct posture of
Salat poses. Our current project is an Artificial Intelligence (AI) based solution for
correct posture detection which can be used for all ages. As a methodology, we have
evaluated two modes or methods: The MediaPipe approach which employs a multi step detector machine learning pipeline that has been proved to work in our Media
Pipe Hands and Media Pipe Face Mesh products. The pipeline initially locates the
person/pose region-of-interest (ROI) within the frame using a detector. Using the ROI cropped frame as input, the tracker then Forecast the pose landmarks and division mask
in between the ROI and another method is Convolutional Neural Network (CNN) both
ways have given us 99 to 100 % accuracy and promising result with real-time
evaluation.
