Autoencoder-based Models for Breast Cancer Detection

This repository explores the use of autoencoders for breast cancer detection using ultrasound image data.

Data

The data used in this study is the Breast Ultrasound Images dataset from Kaggle: Link to dataset on Kaggle.

Autoencoder Models

Three different autoencoder models are implemented:

1. Linear Autoencoder: A basic autoencoder with Tanh activation functions and dense layers as encoder and decoder.

   

2. Convolutional Autoencoder (CNN Autoencoder): A convolutional autoencoder that utilizes convolutional and pooling layers to extract features from the image data.

   

Feel free to use the notebook and contribute to the project!

I encourage you to explore the code, make modifications, and share your findings.