SKINai

• 

  SKINai_application

• 

  SKINai_problem

• 

  SKINai_solution

Inspiration

Due to a lack of educational materials on People of Colour, dermatologists are unable to diagnose PoC adequately. Also, modern solutions using AI also seem to show systematic biases - while being helpful to diagnose diseases for lighter skinned people, the results for darker skin tones cannot convince.

What it does

Skin ai is an app that allows dermatologists to capture and classify images of skin diseases in dark-skinned people and then upload them to the app to make them available to researchers or other physicians. Doctors also have the possibility to view sample images of different skin diseases for different skin types. After enough pictures of skin diseases in dark-skinned people have been collected. Different AI models can be trained to work on darker skin types as well. Which then can be implemented in the app.

How we built it

We developed the idea by researching the problem and searching for possible existing solutions. As we discovered that no specific solution exists right now, we built a prototype showing the basic functionality of the needed application for data collection, education and diagnosing with ai. To do so we used the software Figma.

Challenges we ran into

One challenge was to specify the problem itself as a lot of research is needed. Another challenge was to create the prototype in such a short time as we worked with Figma for the first time. As there are not many similar applications already on the market, we had to make many decisions to create a worthy prototype. Also, it was difficult to know what we actually need to do in an idea hackathon. Therefore we decided to focus on the data collection an application instead of creating an example ai (which is not possible due to the lack of data in the moment).

Accomplishments that we're proud of

We are proud of getting the presentation and prototype done in time and to come up with an idea in such a short time. You can watch our presentation on youtube and test our prototype afterwards yourself.

What we learned

We learned that there seems to be many groups of people, which are not supported in the best way possible in our current health system and that this really is an issue, which should be focused more. We also learned - for ourselves - that setting a limited amount of time for an idea development can really help to come up with something usable.

What's next for SKINai

To get a head start on the data collection, we need to work with a health organisation that is kind enough to provide us with labeled data. After that we analyse this data with the existing models for lighter skin tones to be able to get first results via a scan.

Built With

• css
• css3
• figma
• google
• google-docs