Description as a Tweet:

DermSafe is an Android application intended to provide assistance to people with Skin Cancer in identifying and treating their ailments ASAP. This is a team project consisting of Kushagra Srivastava, Nikhil Jain, Nhan Ton, and Rebecca Wang.

Inspiration:

We wanted to create a health-related project that would utilize ML/AI as well as concepts such as object detection to assist users in identifying various skin-related ailments early on and act upon them. Often, severe skin-related diseases go un-noticed and are not treated upon until it's too late; this can be even fatal to the person's life in most cases. The latest figures according to the Skin Cancer Foundation suggest that more than 15,000 people die yearly due to skin cancer. People can reduce the risk of some types of skin cancer by up to 78% just by knowing of it in the disease's early stage. Therefore, we wanted to create a solution that would be easily accessible by the user and give relevant feedback and resources as soon as possible.

What it does:

The main purpose of DermSafe is to detect and identify what kind of skin-disease a person has based on the pictures they upload to the app. The app uses its own database to compare the user's uploaded pictures and predict the ailment that the user may have.

Users can either use their phone's camera to click pictures and/or upload photos of their skin to know if they have skin cancer and of what type. Users also receive access to a wide variety of resources and contact information that would help them obtain the relevant treatment(s) and keep them well informed about their ailments and preventive measures.

How we built it:

DermSafe was mainly built on Android Studio through a combination of Java, Tensorflow, Keras, and Python. The initial project was open sourced from GitHub: a very basic skin cancer detector (cited in the README file of the Project). Our team repaired the project, and used it as our own canvas, heavily modified it, as well as added various features of our own to compliment and extend upon the original application.

Technologies we used:

Java
Python
AI/Machine Learning
Misc

Challenges we ran into:

Some of the challenges we came across were:

1) Learning how Android Studio works (+ using AVDs for testing)
2) Grasping the Syntax of various different languages out of our experiences: initializing Gradle, working with Java, Tensorflow, etc.
3) Dealing with frequent App Crashes
4) Learning how to work with UI/UX elements
5) Using git effectively across the team.

Accomplishments we're proud of:

As a team consisting of mainly beginners, the biggest thing that we are honestly proud of is getting such an ambitious project done in such short amount of time. Ranging from learning Android Studio, programming in various languages, to using git to collaborate, this was a completely new experience for many in the team.

There were many moments during the duration of the hack where we thought that we may not be able to get this done, and there were many ideas that we were not able to implement (such as integrating Google Maps to direct patients to medical professionals), but we were somehow able to push through and create a completely working app; as well as make the most out of this entire event and the workshops it had to offer.

What we've learned:

Some of the things that we learnt during this project were:

1) Using Android Studio and its tools effectively
2) Creating Java activities in Android, and linking various activities together to execute a task.
3) Working with UI/UX elements
4) Working with Gradle
5) Debugging build errors on Android Studio
6) Using git to work more effectively: especially merging, reverting commits and solving issues

What's next:

Here are some things we intend to incorporate into the project in the future:

1) Expanding the database of the project and detect/provide resources for all types of skin diseases, not just skin cancer.

2) Google Maps Integration to direct Users to the nearest Hospital in case of requirement.

3) Emergency Contact Feature to notify the friends and family of the User in case of emergencies.

4) Incorporating a more fluid UI across the entire app, as well as porting it over to various other operating systems.

5) Including a User-Account setup through which users can save their data on the cloud.

Built with:

Android Studio (and bundled tools: mainly AVD for testing), Java, Python, Tensorflow, Keras, Git/GitHub

DermSafe