[2023 OpenSW Developer Contest Main Round Progress] Non-Face-to-Face Exercise Learning Platform, "MotuS"
This project aims to develop a non-face-to-face exercise learning platform.
Anyone can register a guide video for a specific exercise on the platform, and learners can access the registered exercise programs. Through AI's similarity measurement, users can check the accuracy of their exercise performance.
The core value of the project is that it is not limited to a specific domain, and if there is an exercise performance video, anyone can register and learn from the guide.
| BE | BE | FE | FE | AI |
|---|---|---|---|---|
 |
 |
 |
 |
 |
| DongHun Lee | JuYoung Park | SoHyun Oh | KyungJae kim | InSung Park |
| Hallym University | Hallym University | Hallym University | Hallym University | Hallym University |
| Computer Engineering 4th | Computer Engineering 3th | Computer Engineering 4th | Computer Engineering 2th | Artificial Intelligence 2th |
| @Dong Heon Lee | @JuYoung Park | @osohyun0224 | @PortalCube | @InSung Bahk |
- Non-face-to-face exercise learning platform, MotuS open source video KOR
- Non-face-to-face exercise learning platform, MotuS open source video ENG
- Videos currently registered in the service are displayed with their titles, descriptions, thumbnail images, and tags.
- When a user selects a video, a preview of the video along with the title, description, and tags are shown, along with a button to move to the course page.
- Users can search by video title and tags. Tags allow users to select specific postures or body parts.
- Users can select a video from their computer to upload to the server.
- You can check whether the AI server receives the skeleton data of the uploaded video properly.
- Users can enter the title, description, category, and posture of the video.
- The video that the user wants to take is played for them to see.
- The video is played again for users to follow along while performing the exercise.
- If the user wishes, they can repeatedly view the video before the measurement.
- When the measurement starts, the user's screen is recorded while the video is playing.
- When the video ends, the recording stops and is sent to the server.
- The user's video is analyzed on the AI server, and the similarity to the original video is calculated and displayed.
- Depending on the similarity score, the performance results are displayed to the user as Perfect, Great, Good, or Bad.
| 🫱🏻🫲🏻 H-FIVE Team | 📚 Tech Stacks |
|---|---|
| Collaboration Tools | Github Slack Notion |
| 🧑🏻💻 FrontEnd Team | 📚 Tech Stacks |
|---|---|
| Dev Packages | React Redux Axios |
| Development Support Tools | ESLint Prettier Vercel Vite |
| License | MotuS Frontend License |
| 🧑🏻💻 BackEnd Team | 📚 Tech Stacks |
|---|---|
| Environment | InteliJ Postman Git Action Git Gradle Raspberry Pi 4B |
| Development | Spring-Boot Java NCP(Naver Cloud Platform) MariaDB |
| Dependencies | QueryDsl Spring-Data-JPA Spring Security JWT(Json Web Token) Lombok Validation Gson |
| License | MotuS Backend License |
| 🧑🏻💻 AI Team | 📚 Tech Stacks |
|---|---|
| Frameworks Server | FastAPI , unvicron, mysql |
| Frameworks AI | Pytorch, Numpy |
| Frameworks Pre-processing | Scikit-learn, Scikit-video |
| Environment | Macbook Pro M1 GTX 2070 Super Visual Studio Code Postman |
| License | MotuS AI License |
