Skin Sense: A Skin Cancer Detection App

Problem Statement

Skin Cancer is the most common cancer and is often ignored by people at an early stage. An estimated 178,560 cases of melanoma will be diagnosed in the U.S. in 2018. If melanoma is recognized and treated early, it is almost always curable, but if it is not, the cancer can advance and spread to other parts of the body, where it becomes hard to treat and can be fatal.

Solution

Melanoma, if treated early, could be cured. Our application, with 84.7% accuracy is capable enough to classify Melanoma skin cancer even at an early stage remotly. One can easily classify the type of cancer as Melanoma or not and can take actions accordingly.

Technology Stack

• TensorFlow - Google's DeepLearning Framework

• Model Architecture Inception v3 (mobilenet_100_224)

• Android Studio

• Google Maps API

• Firebase - Google's Mobile and Web App Development Framework

Features

• Offline data storage for statistical analysis and expertise consultancy

• Offline Classification

• Google authentication with email verification

• In app Hospital finder

Screenshots

Special Mentions

Stood first among the 70 teams in Hitsathon V2.0 (Hindustan University Hackathon)

Team Members:

Team Scary Bugs:

• Tanay Toshniwal
• Arnab Ray
• Ritvik Jain
• Rajnish Raj