DiseaseDetector

• 

  mode training

• 

  line graph

• 

  image example

• 

  output prediction

Inspiration

The recent COVID-19 pandemic proved to us that what a disease can do if no one knows what is it. We could not guarantee which disease it was and hence a proper cure could not be administered.

What it does

Our model predicts on the basis of image scans whether a person is infected by a particular disease or not

How we built it

We created this model using convoluted neural network deep learning technique . We used the optimizer adam.

Challenges we ran into

We did not get the required accuracy at the first try, then we ran into overfitting the model, after which we increased the number of layers to get the required accuracy.

Accomplishments that we're proud of

Training accuracy of 99.46% was achieved and we achieved a testing accuracy of 99.19%. We compared to other researcher paper and our model achieved much higher accuracy than all the others.

What we learned

Using deep learning techniques we can easily get a desired accuracy if the model is trained with the right amount of images

What's next for DiseaseDetector

We still want to increase the accuracy, as this is a project based on the medical field and thus the error line is bare minimum and thus would use transfer learning.

Built With

• excel
• google-colab
• google-drive
• python
• sql
• visual-studio