/Merkel-Cell-Carcinoma-detection-

Employs deep learning techniques for the precise classification of diverse skin conditions specifically distinguishing between cancerous and non-cancerous states through the evaluation of skin image

Primary LanguageMATLAB

Medical Image Classification Using Deep Learning

This repository contains a MATLAB script for classifying medical images using various deep learning algorithms, including CNN, DNN, RNN, and ResNet-50. The script uses MATLAB's Deep Learning Toolbox and Image Processing Toolbox to preprocess the images, build and train the networks, and evaluate their performance.

Features

  • User Interface for Image Selection: Allows the user to select an image file using a graphical interface.
  • Data Preprocessing: Handles image data loading, resizing, and augmentation.
  • Convolutional Neural Network (CNN): Implements a CNN with multiple convolutional and pooling layers, followed by fully connected layers for image classification.
  • Deep Neural Network (DNN): Implements a simple DNN with fully connected layers and ReLU activations for classification.
  • Recurrent Neural Network (RNN): Utilizes a sequence input and LSTM layers for image classification.
  • ResNet-50: Implements a simplified version of ResNet-50 architecture for more accurate image classification.
  • Performance Evaluation: Evaluates and prints the accuracy of each network on the training and validation datasets.
  • Visualization: Displays the input image and its predicted class using each model.

Output

https://docs.google.com/document/d/1FjHNc9Y3fqSkxrKUHkCUh87uzkEtFDopJQLMx0vUrfo/edit?usp=sharing