AI-Powered Skin Scanner

The AI-Powered Skincare Scanner is a smart device that uses artificial intelligence and computer vision to analyze facial skin conditions such as acne, acne scars, redness, wrinkles, and hyperpigmentation.

It is designed for individuals seeking personalized skincare, as well as professionals like dermatologists and beauty retailers.

The scanner provides real-time analysis, and delivers recommendations via web/mobile interface.

Supplies

1. Raspberry Pi (for AI processing and control)
2. Laptop (for model execution and display)
3. USB Camera (mirrorless preferred) (for facial image capture)
4. LCD Display (for user feedback and instructions)
5. LED Matrix/LED Strip (for visual indicators)
6. Ethernet Cable (for Raspberry Pi connection)
7. Dummy battery (for the constant camera power)
8. Custom Enclosure Materials:
9. 6mm multiplex wood or plywood
10. Black acrylic paint or spray paint (with white primer)
11. Double-sided tape
12. Black fabric (for light control)
13. Tools:
14. Laser cutter (for precise housing)
15. Drill (for cable holes)
16. Screwdriver and glue
17. Software:
18. Computer vision framework (e.g., YOLO, Roboflow)
19. Dashboard/visualization software (web or local app)
20. Data annotation tools
21. Bill of Materials:
22. Mirrorless camera:
23. Price depends on camera choice
24. Elgato Cam Link 4K:
25. ~100 EUR
26. Dummy battery (for camera):
27. ~20 EUR
28. Raspberry Pi 5:
29. ~120 EUR
30. Freenove Project Kit:
31. ~50 EUR
32. LED strip:
33. ~20 EUR
34. Multiplex wood (6mm, for enclosure):
35. ~70–80 EUR (price depends on store)
36. Black acrylic paint:
37. ~10 EUR

1. Sketch the enclosure: Create side, front, and top views using design software such as Adobe Illustrator or Fusion 36023.
2. Cut the panels: Use a laser cutter to cut all enclosure elements from 6mm multiplex wood.
3. Assemble the box: Glue or screw the panels together, ensuring a snug fit for all components.
4. Prepare for lighting: Install an LED strip inside the enclosure. Use black fabric to cover excess light and prevent glare.
5. Drill holes: Create small holes for cables and ventilation.
6. Finish: Paint the enclosure black for a professional look (use white primer first if needed).

1. Mount the Raspberry Pi: Secure the Raspberry Pi inside the designated compartment.
2. Attach the camera: Place the USB camera in its holder at the back of the enclosure.
3. Install the LCD display: Mount the LCD display at the back for user interaction.
4. Connect the LED matrix: Attach the LED matrix or strip for visual feedback.
5. Power up: Connect all components to a power supply.

1. Collect and prepare data:
2. Use publicly available datasets (e.g., Roboflow Facial Skin Dataset, Nexdata Human Facial Skin Defects Dataset).
3. Supplement with your own labeled images for improved accuracy.
4. Preprocess images: normalize, augment, and annotate as needed.
5. Train the AI model:
6. Use a computer vision model (e.g., YOLO) for segmentation and detection.
7. Train the model to recognize skin conditions: acne, acne scars, redness, wrinkles, hyperpigmentation.
8. Develop the dashboard:
9. Create a real-time visualization interface (web or local app) for displaying results, confidence scores, and recommendations.
10. Ensure bounding boxes and segmentation masks are visible for affected areas.
11. Integrate with hardware:
12. Connect the camera and LCD display to the Raspberry Pi.
13. Enable real-time image capture, analysis, and display.

1. User testing: Invite volunteers to test the scanner during project weeks and collect feedback.
2. Refine the model: Adjust the AI model based on user feedback and new data.
3. Optimize the enclosure: Make any necessary adjustments to improve usability and aesthetics.

1. Add more skin condition categories for enhanced analysis.
2. Expand product recommendations based on detected conditions.
3. Integrate mobile/web notifications for remote access and tracking.

With this guide, you can build your own AI-Powered Skincare Scanner using accessible hardware and open-source software. The project is ideal for makers interested in AI, computer vision, and personalized health technology.