Kaizen: Computer Vision Based Interactive Karate Training Platform
Authors:
- 1 November 2022
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Abstract
All types of martial arts consist of several forms of combat used in self-defense, which are deeply rooted in many countries. Of all the martial art types, karate is considered the most well-known out of them all. Due to the pandemic situation in Sri Lanka, karate enthusiasts have lost the opportunity to train in a well-guided environment. As a result, even though virtual training came into play, it has continuously proved its ineffectiveness in evaluating the performance and accuracy of the trainees. The main objective of this proposed system is to virtualize the processes of a physical karate dojo. Kaizen - A Computer Vision-Based Interactive Karate Training Platform is a web-based application that functions as a virtual instructor. The proposed system consists of two main core components for Training and Assessments. The karate training component evaluates the techniques against a set of predefined joint angles. The BlazePose model is used for keypoint detection, and Analytic Geometry is used to extract joint angles. It is also integrated with Amazon Polly, a Deep Learning-based Text-To-Speech (TTS) service to produce real-time audio feedback. The assessment component has the capability to evaluate the trainees through a built-in Smart Evaluator based on a Recurrent Neural Network (RNN). Additionally, the capability to manage the assessments supports the instructors in conducting all the assessments virtually, overcoming the barriers in physical training.Keywords
This publication has 16 references indexed in Scilit:
- Industrial Robot Control by Means of Gestures and Voice Commands in Off-Line and On-Line ModeSensors, 2020
- iKarate: Improving Karate KataProcedia Computer Science, 2020
- Recognizing hand gestures for controlling home appliances with mobile sensors2019
- Systematic literature review of hand gestures used in human computer interaction interfacesInternational Journal of Human-Computer Studies, 2019
- Artificial Neural Network to Detect Human Hand Gestures for a Robotic Arm Control2019
- Voice-based Road Navigation System Using Natural Language Processing (NLP)2018
- Modeling Augmentative Communication with Amazon Lex and Polly2018
- Yoga posture recognition by detecting human joint points in real time using microsoft kinect2017
- Recording of Occurrences Through Image Processing in Taekwondo Training: First Insights2017
- Human gesture recognition using Kinect camera2012