PHODO : Face Tracking System
Contributors & Oragnization
Contributors
- Keunju Song
- Yechan Yun
- HyeKyung Kim
- SoHee Park
- JungYoong Min
- Su Min Kim (adviser)
Oragnization
- Tech University of Korea LINC+
- 20th Industrial Technology Exhibition, Tech University of Korea
Overview
PHODO is face tracking system that consisted of the main body and mobile app.
PHODO's face tracking technology is based on Machine Learning.
From this system, we can anticipate applications in Entertainment and Security field with facial similarity.
Paper about PHODO is updated. [Click]
See more information about PHODO : https://yyc9920.github.io/phodo.github.io/index.html
System Architecture
System Block Diagram
H/W
- Raspberry Pi Zero W (MCU)
- Stepping Motors (2EA)
- Servo Motor
- LiPo battery
- Rechargeable 5V Lipo USB Boost - Adafruit
- Ultrasonic Sensor - HC-SR04
- Tripod
- Camera Rail
Application
PHODO
- Supporting basic camera function
- Tracking Faces
- Supporting Buletooth that links to the main body
3D Modeling
- Designed by JungYoong Min
Function of the System
- Multiple Face Tracking: When there are more than one face on the screen, tracking the coordinates of the faces by averaging operations.
- User's Face Customizing: PHODO can recognize the face individually, so it can tracking just one person that user want.
- Handsfree Capture: PHODO have voice recognition to support the handsfree capture. Setting voice language is Korean.
- 3-Div Optimal Composition: PHODO can take the photo automatically with optimal composition by controling face coordinates. Face position is established Left-Center-Right.
Result
- PHODO main body
- PHODO Application
- PHODO Face Tracking action
Development Environment
- Android Studio IDE (Java, C++, Kotlin)
- Raspberry Pi Terminal (Python)
Reference
- Face Recognition : https://github.com/pillarpond/face-recognizer-android
- CameraView Library : https://github.com/natario1/CameraView
- Real-Time Face Detection on Android with ML Kit and CameraView : https://heartbeat.fritz.ai/building-a-real-time-face-detector-in-android-with-ml-kit-f930eb7b36d9