FitBeat: Intelligent Music Player Based on Real-Time Workout Monitoring
FitBeat: Intelligent Music Player Based on Real-Time Workout Monitoring
FitBeat: Intelligent Music Player Based on Real-Time Workout Monitoring
FitBeat: Intelligent Music Player Based on Real-Time Workout Monitoring
Organization:
UC San Diego
Organization:
UC San Diego
Organization:
UC San Diego
Duration:
Sep 2023 – Dec 2023
Duration:
Sep 2023 – Dec 2023
Duration:
Sep 2023 – Dec 2023
Video Demo
Demo 1
Demo 2
Demo 3
Overview
Designed an Android app for Samsung Galaxy Watch 4 that can monitor and send real-time health and motion data to the server.
Implemented a real-time workout detector that can recognize current state, count reps, and train machine learning model with personal data, based on computer vision, Attention-Based LSTM, and Google MediaPipe; realized communication with server and smartwatch.
Developed a machine learning model for offline music categorization using Chunk-CNN-Residual; prototyped a Node.js central server that manages communication with devices, generates music recommendations, and controls music stream on Amazon Echo Dot.
See the Presentation Slides for more details.
Video Demo
Demo 1
Demo 2
Demo 3
Overview
Designed an Android app for Samsung Galaxy Watch 4 that can monitor and send real-time health and motion data to the server.
Implemented a real-time workout detector that can recognize current state, count reps, and train machine learning model with personal data, based on computer vision, Attention-Based LSTM, and Google MediaPipe; realized communication with server and smartwatch.
Developed a machine learning model for offline music categorization using Chunk-CNN-Residual; prototyped a Node.js central server that manages communication with devices, generates music recommendations, and controls music stream on Amazon Echo Dot.
See the Presentation Slides for more details.
Video Demo
Demo 1
Demo 2
Demo 3
Overview
Designed an Android app for Samsung Galaxy Watch 4 that can monitor and send real-time health and motion data to the server.
Implemented a real-time workout detector that can recognize current state, count reps, and train machine learning model with personal data, based on computer vision, Attention-Based LSTM, and Google MediaPipe; realized communication with server and smartwatch.
Developed a machine learning model for offline music categorization using Chunk-CNN-Residual; prototyped a Node.js central server that manages communication with devices, generates music recommendations, and controls music stream on Amazon Echo Dot.
See the Presentation Slides for more details.
Video Demo
Demo 1
Demo 2
Demo 3
Overview
Designed an Android app for Samsung Galaxy Watch 4 that can monitor and send real-time health and motion data to the server.
Implemented a real-time workout detector that can recognize current state, count reps, and train machine learning model with personal data, based on computer vision, Attention-Based LSTM, and Google MediaPipe; realized communication with server and smartwatch.
Developed a machine learning model for offline music categorization using Chunk-CNN-Residual; prototyped a Node.js central server that manages communication with devices, generates music recommendations, and controls music stream on Amazon Echo Dot.
See the Presentation Slides for more details.
