Inspiration
Statistically, 1 out of every 5 people suffer from skin cancer at some point in their lives. Diagnosis of skin cancer is a very long and expensive process, and still gives 83% accuracy. We wanted to solve this problem using artificial intelligence to make this process more efficient and scalable to reach the masses.
What it does
A patient uploads their skin image on our website. That image runs through our trained Convoluted Neural Network (CNN) model, which gives a data-driven decision on whether that patient is suffering from skin cancer.
How we built it
We coded the CNN in Python using pandas, and used react to build the website
Challenges we ran into
1) Exporting the database to GitHub repository 2) improving accuracy of deep learning model 3) exporting the model to embedding it into the website 4) connecting firebase to custom domain (registered from Domain.com)
Accomplishments that we're proud of
1) achieving accuracy of 83% on our test data 2) Built a user-friendly website to function
What we learned
1) Ideating and concluding a project within 24 hours. 2) Collaborating in a team and delegating tasks based on individual strengths 3) Smoothly integrating Deep learning model with front end
What's next for Detecting Skin Cancer using Deep Learning
1) getting more data to be able to train our model better (for better accuracy) 2) making the website more user-friendly
Log in or sign up for Devpost to join the conversation.