DermaScan

Skin Disease Classifier

Please wait for 5 seconds before uploading image

Accepted image format ↪ jpeg or png

Results are not a substitute for professional medical advice.

Predicted Disease

samplepic.jpg
Benign Keratosis: 0.385
Melanocytic Nevi: 0.366
Dermatofibroma: 0.204

SIH 2023 PPT

How it Works

Image Input

A high-resolution image capturing the entirety of the affected skin region is uploaded and then transmitted to the backend for further processing.

Image Processing

The image is subsequently processed to ensure compatibility with the MobileNet V2 model.

Image Classification

Then the image is fed into the model, which then predicts top three most likely skin conditions from a list that includes actinic keratosis (AK), basal cell carcinoma (BCC), benign keratosis (BK), dermatofibroma (DF), melanocytic nevi (NV), melanoma (MEL), and vascular skin lesions (VASC).

Results Display

After the MobileNet V2 model generates predictions, they are transmitted from the backend and presented to the end user on the frontend.

Technologies Used

This project leverages Python, Keras, TensorFlow, scikit-learn, HTML5, CSS, JS, and Bootstrap, hosted on GitHub, to create a dynamic and interactive website.