Carcinoma - Skin Cancer Detection

Skin Cancer Detector - to detect skin cancers without going through mentally traumatizing image(s)

Comment

Inspiration

To provide an easy to use skin cancer detection app for patients as well as doctors. (Skin Cancer can be easily cured if detected at early stages).

What it does

It detects Melanoma and 7 other types of skin conditions with 80% accuracy.

How I built it

The model was trained using Tensorflow and Keras. The frontend is designed using HTML, CSS & Javascript. We used bootstrap library for styling. Ajax was used for updating the web content. Backend is designed using Flask.

Challenges I ran into

Integration issues, Deployment issues, Training issues (Lack of Infrastructure)

Accomplishments that I'm proud of

Developed a user friendly web app which can be used by everyone.

What I learned

Make web app around deep learning models.

What's next for Carcinoma - Skin Cancer Detection

We are planning to extend this project into a full-fledged medical symptoms solution. We would be using infermedica API to build a chatbot to detect the possibility of underlying disease by understanding evaluating symptoms of the patient. The web app would be further extended to android and IOS apps to increase usability and ease of use. Handheld skin cancer disease detector can also be made using Raspberry Pi.

Share this project:

Updates