Runner's Movement Recognition | Machine Learning
Runner's Machine Learning is real-time movement recognition project. The project is resulted in classification of walking, running and squats with 96% success on our dataset. The project is continuum of Runner's Android Application.
Tools & Techs Used
Software: Python (keras, tensorflow, sklearn, matplotlib), Brackets, Anaconda, Socket.io, WebStorm
Algorithms: Long Short Term Memory (LSTM) and Random Forest (RF). Support Vector Machines (SVM) are used before RF, but RF has chosen over SVM for its running speed.
Hardware: Bluno, soft sensors and microcontrollers
Check it out here: https://youtu.be/ZryvdvgoNlI
Technical Demo: https://youtu.be/O2oVdo0vo8A
This project is selected among the best projects of the year. Decision is made by the professors of our faculty after the presentations (here is mine) and then we've prepared project posters to present it to the company owners. You can access the project poster here.
I have developed the app as my Computer Engineering graduation project. I've developed an android app to show movement data to the user in real-time. I wanted to improve myself on artificial intelligence and contribute to the society. This idea was proposed to me by my advisor and just after that moment I have started working on it. There's a long way to walk, though I enjoyed working on the project and learning about artificial intelligence and also python.
Unfortunately, the application is developed for a private project held by my faculty (Computer Engineering, Istanbul Technical University, Istanbul/Turkey). I have created this repository to share the insights and my work during my graduation project.