DermaBuzz

• 

  DermaBuzz

Inspiration

With skin cancer becoming one of the most common cancers globally, we were inspired to create an accessible, AI-powered tool to assist with early detection. We believe that by empowering people with easy-to-use technology, we can help catch potential signs of skin cancer before it progresses, especially in regions with limited access to dermatologists.

What it does

SkinSnap is an app that scans images of skin lesions and uses machine learning to determine the likelihood of skin cancer. Users can upload a photo of a suspicious mole or spot, and the app instantly analyzes the image, providing an assessment of whether further medical consultation is recommended.

How we built it

We used a convolutional neural network (CNN) model trained on skin cancer image datasets like ISIC and HAM10000 to classify images. The front end of the app was built using React for an intuitive user experience, while the back end, powered by Flask, processes the uploaded images and performs the AI-based analysis.

Challenges we ran into

One of the biggest challenges was dealing with the imbalanced dataset, where healthy skin images outnumbered cancerous ones. Balancing the training process and avoiding overfitting while maintaining high accuracy was tricky. We also had to carefully design the user interface to ensure that the app remained simple yet functional for non-technical users.

Accomplishments that we're proud of

We’re proud of creating an intuitive tool that simplifies the complex process of skin cancer detection. We managed to achieve a high accuracy rate using transfer learning with pre-trained models, and we optimized the model to run efficiently on consumer devices. Additionally, we developed a sleek user interface that allows users to easily upload and scan images.

What we learned

Through this project, we gained a deeper understanding of medical image processing and the ethical considerations involved in deploying healthcare AI. We also learned how to balance AI accuracy with user experience, ensuring the app provides valuable insights without overwhelming users.

What's next for DermaBuzz

We plan to improve the app by expanding its dataset to include a wider range of skin tones and conditions. We’re also working on adding more features, such as real-time consultation with dermatologists and tracking changes in skin over time. Long-term, we aim to launch SkinSnap as a free tool in partnership with healthcare providers to reach underserved communities globally.

Built With

• amazon-web-services
• are
• as
• cloud
• employed
• ensuring-a-seamless-and-engaging-experience.-to-enhance-image-processing
• for
• google
• leveraging-frameworks-like-tensorflow-and-keras-for-model-development.-flask-serves-as-the-backbone-for-the-api
• managing-requests-and-responses-efficiently.-the-user-interface-is-crafted-with-react
• opencv-is-utilized-for-preprocessing-tasks
• or
• platform
• scalable
• services
• such
• technology-stack-skinsnap-is-built-using-a-combination-of-robust-technologies.-the-backend-is-powered-by-python
• while