SelfDerm

Making Dermatologist Accessible

About

       SelfDerm is a mobile application which helps in diagnosing skin diseases. This app could be used by skin-related patients for tracking their moles and dermatologists for being confident about their diagnosis. This app is a small step towards better skin care & a study shows that misdiagnosis caused a lot of deaths so this app could help to reduce the misdiganosis made by dermatologists. The app is using Deep learning to diagnose skin diseases of 18 Categories. The app is built in Flutter & Firebase with a lot of features & Great UI UX. This app requires a lot of improvements related to AI and app end as well.

Problem

       Diagnosing skin disease is an important part of dermatologists job. When dermatologists are diagnosing skin diseases there is a lot of chances of misdiagnosing which could leads to some several consequences like if a skin mole is going to become a dangerous cancer and it's diagnosed as a normal skin mole then it could leads to the death of patient. There is another problem related to that which shortage of dermatologists so in that normal people needs something where they can diagnose their skin moles and take quick actions.

Solution

       The Artificial intelligence is giving us the solution to this problem. I have diagnosed skin diseases using CNN: a deep learning algorithm. I have collected the dataset of skin disease images from different open source platforms and combined them according to my requirements. I have used the Google colab, python, tensorflow, Keras, Opencv and other tools to train my AI model. Once the AI model is trained successfully then I have convert that into mobile version by using TfLite and made an mobile application in Flutter for end user to use my AI model. The app also contains a lot of other features which are facilitating the end user.

Description

       It's my portfolio which you are seeing right now. There is also an admin panel which you can visit by using the Credentials & Demo link.

Outcome

       It's my portfolio which you are seeing right now. There is also an admin panel which you can visit by using the Credentials & Demo link.

Interesting Features

• Identifying the given picture is the picture of skin mole or some random picture
• Diagnosing skin Disease from single picture of mole
• A CaseLog where user can view all of his cases( Each diagnosis made by AI is considered as a single case)
• A simple Authentication using phone number

What I have Learned

• Making Great UI UX

Tools & Technologies