Skip to content

Cross-platform smartphone app capable of detecting skin cancer lesions using Computer Vision.

Notifications You must be signed in to change notification settings

msthoma/LesionDetector

Repository files navigation

app build

Lesion detector app

A smartphone app built with Flutter, capable of detecting skin cancer lesions using Computer Vision. To do so, the app incorporates a Convolutional Neural Network trained with the HAM10000 dataset, which consists of "a large collection of multi-source dermatoscopic images of common pigmented skin lesions".

Convolutional Neural Network information

The CNN used was trained using TensorFlow and Python, as explained in detail in this Github repository and this Google Colab notebook.

App information

The app is built with the Flutter framework, and so it works on both Android and iOS. Most of the app's functionality is implemented in this file: /lib/main.dart.

Releases

No releases published

Packages

No packages published

Languages