Introduction

Say hi to Incognito, the app which normalizes healthcare, in a great way. Incognito aims to diagnose Parkinson’s Disease in 3 intuitive steps: draw a spiral, point, and click (a picture). Parkinson’s disease is the second most common age-related neurodegenerative disorder after Alzheimer’s disease, but what makes it so unique is its external diagnosability: no longer do we need fancy MRI or PET scans to diagnose Parkinson’s Disease. Parkinson’s Disease is only getting worse, increasing from 41/100,000 people to over 190,000/people, but here’s the catch: we’ve only looked at people who are able to visit doctors, and people who have the healthcare means to get diagnosed. What about the 46.5 million Americans who lack healthcare, or the millions more worldwide who do not have access to good healthcare. Parkinson’s is scary: a patient of this disease will no longer have a normal social life or have a job, therefore making them incapable of paying for medical expenses, and will severely hurt their financial condition. In the United States alone, 60,000 people are diagnosed with Parkinson’s every year, however, this fails to reflect the over 40 million Americans living under the radar, and every Parkinson’s statistic fails to represent millions of people who are Incognito to the healthcare system. Incognito brings light to those who otherwise would not be able to attain healthcare.

Inspiration

The Incognito team was inspired by every day experiences, whether it was being too lazy to visit a doctor for any doubts we had about our health, or the fact that we were privileged enough to visit a doctor whenever necessary. We wanted to do something that could change lives, and we decided that healthcare would be the best way to do so, because through Incognito, we will be able to save lives, by diagnosing critical diseases in their early stages. Furthermore, we were inspired by family members who suffered from Parkinson's Disease... maybe if they had access to good healthcare, their lives would have been saved. Maybe, if they had been diagnosed earlier, they could have lived a more comfortable life. Thats why we're here: we believe that everyone deserves a chance at living, no matter what their background is.

What it does

Incognito is intuitive: draw a spiral, take a picture, and upload it to our webapp to diagnose whether you have Parkinson's, in under a minute. The Incognito team loves simple things, whether it is our simple UI, our easy-to-follow instructions, or the convenience that comes with a web app! In fact, Incognito only requires a phone camera. That's it. No MRI scans, no PET scans, and definitely no microscopes. The user simply draws a spiral, clicks a picture, and gets diagnosed, in under a minute. Instead of taking a long trip to the hospital and waiting amongst people with the flu, we bring the hospital to you. Instead of paying an absurd amount of money to get scanned, we do the fancy image analysis for you. Incognito is simple to use, for anyone and everyone. Whats more? We compete with the pros when it comes to accuracy, but the difference is, Incognito is easy, free, and time-friendly (we of all people know the importance of efficiency)!

How we built it

The Incognito team utilized Convolutional Neural Networks, CNN's provide the highest accuracy when it comes to image detection. We developed an efficient preprocessor to improve the results of our model, including tools such as OpenCV. In fact, we put heavy emphasis on optimizing our code, whether it was through gray scaling images or using the smallest sized images as possible. A lot of work was done on the model to achieve high accuracy, and we've perfected it to our best ability. Furthermore, we utilized bootstrap and css for frontend development, as it is intuitive and provides the most refined results. Much of the frontend was developed through templates that we found visually pleasing. Django was used for further backend processing.

Challenges we ran into

The Incognito team ran into numerous challenges, whether it was timezone differences or technical difficulties. Our ML developers were having issues with their pickle files, which took much googling to fix. Because the Incognito team took a long time to find an idea we were passionate about, it became difficult to do so many tasks in such little time. Our frontend and backend developers worked tirelessly to have a working and visually pleasing User Interface, which was very rewarding. We had to decrease the size of our model in order to host it, which we found incredibly challenging, due to the fact that we wanted to maintain our accuracy. With a lot of tweaking and shifting, we were able to gain maximum accuracy for a model of its size. We have also faced numerous issues when it comes to hosting the web app, which is something we are currently working on (after all, we want our users to have the best time possible). At Incognito, time is our biggest challenge, whether it is beating Parkinson's, or developing our app in under 24 hours.

Accomplishments that we are proud of

We are proud of our entire web app, whether it is our important cause, or our app which was completed in such a small timeframe. We are proud of the fact that when we do launch our app, it will help millions who do not have access to healthcare (and maybe inspire other developers to tackle the same challenge). We are proud that we were able to locally host our web app, which was definitely unexpected, and we are most proud of being able to achieve a high accuracy on our Neural Network. On a less important note, we are proud that our website doesn't look like a virus (We're looking at you, Oracle). Even less importantly, we think our name is pretty slick.

While we are proud of every single part of our app, we know that it is not perfect. We would like to make ourselves and our users prouder, whether it is through improving our UI, increasing our model's accuracy, detecting more diseases, and maybe, just maybe actually hosting our app for the public (just kidding, we'll definitely be doing that).

What we learned

The Incognito team learned a lot. We learned how to develop a fully functioning web app in under 24 hours, which is something none of us have done before. Our ML, frontend, and backend developers had to go into the nitty gritty details about each aspect, therefore solidifying our understanding of all three areas further. We also learned how to wake up before 2pm, but thats a discussion for another time.

What's next for Incognito

The Incognito team has already started expanding, as you are reading this. We plan to develop a large team of capable members to create models for multiple diseases, so you can import any photo, and get yourself diagnosed. It will become a hospital on a 5.6 inch screen (yes, we searched up the average smartphone size). We also plan on improving our model accuracy, and of course, improving our UI and hosting our app. Incognito fully plans to fulfill its name, and cover people with all kinds of conditions who may be incognito to the healthcare system. Incognito will help millions, you can take our word for it.

Thanks for reading (you deserve a cookie, but due to a special virus, we can't give it to you)!

Instructions

  1. Draw a spiral on a piece of paper

  2. Take an image of your spiral

  3. Go to http://amazingbuilder123456.pythonanywhere.com/

  4. Upload the spiral

  5. Enter your email

  6. Submit

  7. Check your email for results!

Side Notes

Accuracy for the model is not too high because we chose to compromise accuracy to have the model fit into the website. In the future we can deploy the model as a microservice and help bring up accuracy.

(Only read if you are a sigma hacks member) After the sigma hacks deadline we were able to integrate the AI model into the website and make it work. You can look at the newly improved website now.

For flare hacks we were able to get the website working and integrate the AI model!

Built With

Share this project:

Updates