Inspiration
EarlyVision AI was inspired by the alarming statistics highlighted in the Medical News Today article, which underscores the increased mortality rates of skin cancer among African Americans. The article emphasizes that skin cancer often goes undetected until it reaches advanced stages in darker skin tones, making early detection crucial. This disparity motivated us to create a tool that empowers individuals within the African American community to take proactive steps in monitoring their skin health, thereby addressing a critical gap in healthcare accessibility and awareness.
What it does
EarlyVision AI is designed to help African Americans detect early signs of skin cancer by analyzing photos of skin anomalies. Users can upload images of suspicious moles, tags, or other skin irregularities, and the application utilizes the Google Gemini LLM to assess the likelihood of cancerous growths. Based on the analysis:
High Probability: Users are advised to consult a medical specialist for further evaluation.
Low Probability: Users receive reassurance that the spot is likely benign, reducing anxiety and unnecessary medical visits.
This dual-response system not only aids in early detection but also provides peace of mind, fostering a proactive approach to skin health within the community.
How we built it
We developed EarlyVision AI using .NET with Razor Pages to create a seamless and responsive web application. The core functionality integrates the Google Gemini LLM API to process and analyze user-uploaded images. The application workflow involves:
Image Upload: Users take a photo of a skin anomaly and upload it to the platform. Analysis: The Google Gemini LLM processes the image, evaluating features that may indicate cancer. Feedback: Based on the analysis, the app provides personalized advice on whether to seek further medical consultation or not. We focused on creating an intuitive user interface that guides users through the process effortlessly while ensuring data privacy and security.
Challenges we ran into
One of the primary challenges was fine-tuning the prompts sent to the Google Gemini LLM to ensure that the application provides appropriate advice without overstepping into giving medical diagnoses. We had to carefully calibrate the AI to recommend seeking professional medical advice when a high probability of cancer was detected, rather than presenting itself as a medical authority. Balancing accuracy and responsible guidance was crucial to maintaining user trust and compliance with medical guidelines.
Accomplishments that we're proud of
Effective Integration: Successfully integrating the Google Gemini LLM API to analyze skin images with high accuracy.
User-Centric Design: Developing an intuitive and accessible interface tailored to the needs of the African American community.
Impact Potential: Creating a tool that addresses a significant health disparity, potentially saving lives through early detection.
What we learned
Throughout the development process, we gained valuable insights into the complexities of leveraging AI for medical applications, particularly the importance of prompt engineering to ensure responsible and accurate outputs. We also deepened our understanding of the unique healthcare challenges faced by the African American community, reinforcing the need for culturally sensitive technological solutions. Additionally, we enhanced our skills in .NET development, API integration, and user experience design.
What's next for Early Vision AI
Moving forward, we plan to expand EarlyVision AI's capabilities by:
Enhancing Accuracy: Continuously improving the AI model's accuracy through more extensive training data and advanced algorithms.
Mobile Integration: Developing a mobile version of the app to increase accessibility and convenience for users on the go.
Medical Partnerships: Collaborating with dermatologists and healthcare providers to validate the tool's effectiveness and integrate it into broader healthcare services.
Community Outreach: Launching educational campaigns to raise awareness about skin cancer risks and the importance of early detection within the African American community.
Log in or sign up for Devpost to join the conversation.