Inspiration
For a better healthcare system for those who have skin issues The healthcare system is generally expensive and people with skin issues more often than not will refuse to get medical attention until it is too late.
What it does
This web-based application takes skin lesion images as input, utilizes deep learning algorithms to classify the skin lesion into one of the 7 categories, which helps with either self diagnosis or professional diagnosis
How I built it
We used 3000 images that are labeled in 7 categories to train the convolutional neural network
Challenges I ran into
The size of the data set was too large for our limited computational power and time constraint. We have to use chunk of the full data set to train the network
Accomplishments that I'm proud of
What I learned
How to build simple convolutional neural network that takes different attributes as input (i.e.: categorical data, numerical data, and image files)
What's next for Yeet Brain
Use more images for each class and more classes of images to further train the model, include more user input options in the front end, such as age input, gender input etc. to better assist diagnosis.
Built With
- css
- django
- html
- javascript
- keras
- python3.7
- sql
- tensorflow
Log in or sign up for Devpost to join the conversation.