Caryio / FaceRecognition

Implementation of real-time face recognition based on CNN on PYNQ Z2

Implementation of real-time face recognition based on CNN on PYNQ Z2

Demo Video

This project was started in 2022.02 and finished in 2022.06.

Content

• Abstract
• Preparation
• Approach
• How to Reproduce
• Thanks

Abstract

Independently implement a real-time face recognition Python project based on CNN convolutional neural network to display and label the captured real-time images and face recognition results on the monitor. Score: 95/100.

Preparation

• PYNQ Z2 board
• USB Webcam
• Monitor
• A computer with Jupyter Notebook

Approach

1. Import all packages needed. Download the bitstream.
2. Configure the video in/out and HDMI in/out.
3. Load and encode the pictures.
4. Detect and recognize the faces.
5. Release the cache and memory.

How to Reproduce

1. Download the file code.ipynb.
2. Turn on the PYNQ Z2 board; USB webcam; monitor. And make sure all components work well.
3. Run the code.

Thanks

• Many, many thanks to my teacher, Prof. Pan.
• Thanks to TA, Dr. Wang and Dr. Huang.