The intelligent beauty mirror project addresses the growing demand for personalized skincare and cosmetic solutions in both consumer and professional markets. Traditional mirrors provide limited functionality, leaving users reliant on manual observation and external skincare advice, which often lacks accuracy and consistency. To bridge this gap, our team designed an innovative smart mirror incorporating advanced image processing, artificial intelligence, and interactive augmented reality features.

This intelligent beauty mirror not only enhances personal skincare routines through real-time skin condition analysis but also provides personalized product recommendations and interactive virtual makeup trials. By targeting professional makeup artists, skincare specialists, beauty enthusiasts, and everyday consumers, the mirror significantly improves the user experience, making beauty routines efficient, interactive, and highly personalized.

The Magic Mirror system integrates real-time skin condition analysis and augmented reality (AR) makeup rendering based on facial landmark detection. The system uses Picamera2 for live video input, MediaPipe for face mesh tracking, OpenCV and Gabor filtering for wrinkle and texture analysis, and Tkinter for graphical user interface design. A voice control module is integrated to support hands-free operation via predefined commands. Additionally, an Automatic Turn ON and OFF Feature was included to control power management and for the convenience of the user.

Detailed explanation of implementation:

The Automatic Turn ON and OFF feature was implemented for convenience and for power consumption management. Our goal for this feature was to reduce the overall power consumption of our project to reduce energy use which would help reduce cost for the user and wear to the components by preventing overheating from prolonged use. This feature would make sure that the device is not being powered when not in use, preventing the user from leaving the main processing unit on accidentally.

For the Automatic Turn ON and OFF Feature we used a combination of PIR sensor, relay switch, and mmWave sensor connected to a Raspberry PI Pico. The PIR sensor was used as a trip wire to see if an object was close enough to the mirror and in front of the mirror. Once the PIR sensor is tripped the Pico would wake up and turn on UART communication to the mmWave sensor. If the mmWave sensor detected a human than the Pico would send a signal high to the relay switch which would connect GPIO 3 and ground of the Raspberry Pi 4. Thr Raspberry Pi 4 has the Howchoo pushbutton software installed, which when GPIO 3 and ground are connected the Raspberry Pi 4 will either turn on or off. You can see the flowchart of this implementation in Appendix B below.

The feature was implemented on the Raspberry Pi Pico in C which is a lower-level processing language that can process much faster and use less resources compared to higher level processing languages such as python and C++. Using the “hardware/sleep” Pico library we can run a sleep command the reduces the power consumption of the Pico from 100 mW to 6.66 mW till the wake pin reads a high signal which is connected to the PIR sensor. When on the mmWave sensor is used to detect if a person is standing in front of the mirror with a 60-degree angle of detection. As long as a person is within the range of the mmWave sensor no signal will be sent to the pulse pin activation the relay and turning the Raspberry Pi off.