SkinSage

Illuminating Paths to Healthier Skin, Brighter Lives.

Comment

Inspiration

Skin disease is one of the most neglected type of disease a simple dandruff is also a type of skin disease. From what I observed and research in India specifically dermatological sector requires a growth rate of 12% annually whereas current is below 10%. It is therefore a concern due to alarming rate of increase in different types of skin disease which can be harmful. In rural areas medical services are very hard to get specifically dermatologists, due to the lack of development and neglection. For just a simple check-up patient has to travel long distance which isn't cost efficient. So, to tackle this I have created SkinSage a webapp which allows user to diagnose their skin disease.

What it does

SkinSage is a skin disease detection webapp built on Django framework. It allows anyone to register an create a account. After that they can easily diagnose their kind of skin disease using the AI trained Deep Learning Model for skin disease detection. As of now my model can detect 8 classes of fatal skin diseases. User can simply upload a decent quality of their skin disease, after clicking on diagnose our DL model will do it's work and classify the disease. And based on the outcome it will give suggestions about what the patient can further do to cure it and precautions for other members of family and patient himself. After diagnosing user can click on book appointment where user based on their location can find good dermatologists near them. They can then proceed to book appointment with the doctor they choose.

How we built it

What I did was:

  • First: The main thing was to build the model the main working part of Skin Sage, I researched on existing technologies and models already built. Very less had used ResNet. I used CNN or better RNN ResNet-101 to train my custom model which gave an unparalled accuracy of 99.43%.
  • Second: Now when the model is done it's time for making an application, and as I have a great hold of in web development I chose to make a webapp. I used Django framework for seamless and secure backend operations for frontend (HTML, CSS, Bootstrap, JavaScript). *Then i integrate my model within my website and made it to a fully functional website ## Challenges we ran into There many challenges I rant into:
  • Data collection: a better quality and quantity dataset is what makes a Deep Learning model better so it was quite a difficult task to collect good dataset.
  • Model Training: making an accurate model was also a difficult task as it takes a lot of time in training due to hardware limitations and then low accuracy and again training is tedious task. But with better data pre-processing a precise result was obtained.
  • Website integration: integrating model to my website was tough task for me as it was something new for me. ## Accomplishments that we're proud of At the end I'm proud of what i was able to achieve and build. A fully functional skin disease detection website with user login database and booking system it was a successful project.

What's next for SkinSage

Future Scope for SkinSage:

  • Making a mobile app for IOS/Android: It is always easier to access a app rather than a website, it is handy and consumers tends to like Apps.
  • Commercial: Making SkinSage open to commercialization, requiring a limited fees for diagnose, then booking appointment(commission from doctors for listing)
  • As for SkinSage making medicinal suggestions adding new parameters to model to precisely and accurately diagnose patients like (skin type, allergies etc.)
Share this project:

Updates