Recall

Inspiration

Many of us have been affected or have seen loved one being affected by many deadly, chronic and incurable diseases. We want to help innovate so that living with diseases like Alzheimers or mental disorders becomes easier for you and your loved ones.

What it does

This is a general doctor's assistant, purposed for helping doctors and patients side by side for the benefit of the patient. This app will have a doctor's interface, where it helps them recommend prescriptions or therapy, accurately classify diseases based on symptoms or images, and track a patient's progress. The patient will have another interface personalized to them, catering to them by providing everything from advice on daily habits to music recommendations.

How we built it

We created our own CNN using Tensorflow to view MRI images and determine what stage of Alzheimer's disease a patient is at. We also created multiple models on ReSpell to recommend medications, therapy, and diagnose other diseases. We connected everything through Flask to make an app.

Challenges we ran into

The TensorFlow model was particularly very hard to train, and it was very tricky to ensure the dimensions of the images were compatible with the model. We tried using data augmentation to help the accuracy of the model and we tried out many different pre-trained models as well. Since we made many different models, connecting everything via Flask was tricky because we had to make multiple API calls.

Accomplishments that we're proud of

We are very proud of our idea and believe that once we implement this app to be effective for doctors dealing with all types of patients, and personalize the app for patients depending on their disease, this app will make a big difference in the health industry.

What we learned

We gained valuable experience in several areas of web development and machine learning throughout this project. On the web development side, we learned how to create a web app from scratch using Python and Flask. We also gained experience with HTML, CSS, and JavaScript by creating the user interface and incorporating it into the app. On the machine learning side, we learned how to build, train, and evaluate a neural network using TensorFlow. We also learned how to preprocess and augment image data to improve the model's accuracy. Additionally, we have gained experience with deployment by hosting your app on a cloud platform like AWS. Overall, this project provided us with a practical, hands-on learning experience that has enhanced your skills in web development, machine learning, and deployment. Two of our members are not technical and are on the medical side, so they learned a lot in terms of machine learning as well.

What's next for Recall

Next, we want to implement some AI generation features. It could be possible to use information about a person with Alzheimers for example and use that information to generate images for them that might remind them of their past as a form of art therapy, or generate music that is similar to music they had growing up. We want to expand this app for all diseases as well, and personalize the interface for patients depending on what disease they have. We also want to fully publish this app so that it can be available to the public.

Built With

• amazon-web-services
• augmentation
• bootstrap
• cnn
• data
• flask
• keras
• python
• respell
• respell-api
• tensorflow
• vscode