Diagnosis AI

With the purpose of helping patients and doctors to detect skin diseases with higher accuracy, Diagnose AI utilizes RESNET AI Deep Learning Network and Transfer Learning to automate the diagnosis.

Comment

Inspiration

Knowing that the medical care cost in the USA is high while the number of doctors cannot meet the need of the patients, we seek to find a solution to create an AI Model to streamline the diagnosing process.

What it does

Using the images of the skin diseases provided, it will use the AI Model to diagnose the skin disease given in the image.

How we built it

Using RESNET AI Deep Learning Neural Network and Transfer Learning with Python, we were able to create a simple website to upload and diagnose the skin diseases.

Challenges we ran into

Since we trained the model with numerous images, the processing used a lot of the GPU Memory Cache, which slows down the diagnosis process to ~2 minutes

Accomplishments that we're proud of

We are now able to detect any skin diseases with 60% accuracy within 2 minutes.

What we learned

We learned how to train an AI Model with Images and dealing with GPU Memory Cache while training the model.

What's next for Diagnosis AI

Diagnosis AI will be more useful for patients on Mobile Device form.

Built With

Share this project:

Updates

Submission history