Swing On

Role

I am a project manager. I communicate and deliver information between the clients and the team. I also make sure the team's progress is on track and all the work satisfy the requirements. I made the block diagram, gantt chart and user interface diagram showing below. 

I also worked on using machine learning algorithm to analyze the data of the body joints. I did the video and image processing, which I split the video and analyze it frame.

User's Experience

The SwingOn smartphone app tracks the motion of the golfers, so they can analyze their swing and receive feedback conveniently on the golf course or driving range. The analysis is performed in real-time on live and pre-recorded videos uploaded from the user’s Photo Library or iCloud Files. The SwingOn app uses pre-trained machine learning models to detect and draw the golfer’s body points. These points are used to compute their centroid and measure their balance. Instant feedback will also be provided after analysis to give the user better insight into improving their swing in the future. SwingOn was created to make golf more accessible for people of all backgrounds and experience levels.

Features

This project is a combination of software development, data analysis, visualization and video/image processing. 

• Software development

  • SwiftUI is used to develop the app

• Data analysis

  • CoreML is used to integrate machine learning models into app. 

  • Pre-trained Deep-learning Neural Network: PoseNet is used to take a processed camera image as the input and outputs body key-points

  • Unsupervised learning algorithm K-mean clustering is used to classify a swing as good or bad

• Video/image processing

  • A video is split into frames 

  • Each frame are feed into the machine learning algorithm.