SkinVue

SkinVue Device

REPORT : Project description: SkinVue is a personalized user-friendly device that detects potentially cancerous moles. It utilizes image recognition (AI) in order to analyze based on factors such as; dimensions, color, and growth. In addition, considering user information for better accuracy, this includes race, age, skin color, and symptoms. These characteristics are then summed and compared to a database in order to determine the potential of this melanoma to be cancerous.

Furthermore, the interface offers continuing monitoring tracking the characteristics of the melanoma. The device can suggest whether you should self-monitor or seek professional medical advice from a doctor.

Inspiration: The development of SkinVae was driven by a personal experience within our team. A team member's melanoma went unnoticed until a family member pointed out its unusual growth. Unfortunately, without knowledge of concerning mole characteristics, there was limited action it could be taken. This situation highlighted two challenges: Limited Visibility, moles on the back are difficult to see regularly, making it hard to track changes. Growth Monitoring: Melanoma can grow over time, but without a precise reference image and consistent measurements, identifying these changes can be challenging.

Inspired by this experience, we envisioned a device to address these limitations. SkinVae would provide regular monitoring, tracking the mole to identify any developments, and analyze with AI to assess potential malignancy, while at the same time creating a simple interface for usage.

System AI image recognition neural network, was trained to recognize the appearance of cancerous melanomas and their specific origin.

The acquisition of an image using the components of ArduCAM Mini with UNO R3 which is located in a casing that provides support in order to be localized in specific areas of the body for analysis.

Challenges? 10:00 am - 6:00 pm Time constraints of completing a 3D model and being given to a secondary provider, [Done with beginner experience]. Pieces collected at 11:00 pm 10:00 am - 3:00 pm Data Collection, Cleaning Data, and Training of AI model was successfully completed. 10:00 am - 5:00 pm ESP8266 to webcam connection complication on recognizing the ArduCAM Mini 3:00 am - 8:00 pm Evaluate, Save weights, and connect to the back end. 5:00 pm -11:00 pm Change system from ESP8266 to Uno R3 complication on recognizing the ArduCAM- Min 11:00 pm -2:00 am Connection with the camera was successful and obtaining an image, completed 2:00 am - :* Sending image to website 8:00 pm -:* Refining and testing

What is next? Further, optimize the user experience with a clear interface and intuitive controls. Enhance image capture with better hardware and image processing directly independent of a monitor. Nevertheless, improve the wearable versatility of the device through more precise and fixed positions including a more efficient way of implementing the COU.