CNN modelling on Chest-Xray data set

Published: 31 March 2026| Version 1 | DOI: 10.17632/4k5gcs2fcs.1
Contributor:
BISWANATH MAHANTY

Description

The CNN modelling exercise here are for classifying "Normal" and "Pneumonia" from chest-Xray data set. The image data itself is not included because of size - but the structured in folder -subfolder, the main folder have subfolders train, test, validation. In each subfolder, there are two folders called "Normal" and "Pneumonia" containing relevant images. The matlab script includes two files (i) using CNN model from scratch (ii) using pretrained network architecture

Files

Steps to reproduce

The two script named CNN_new.m and CNN_pretrained are used to build CNN model based on Chest-Xray image for binary classification. In both the script, the data set for training, validation, and test are repartitioned, the training set is over-sampled for class imbalance, and augmented with rotation, transition. For the pretrained network (CNN_pretrained.m), the input dimension is used to resize the image data store, and last 2 layers are removed with fully-connected, SoftMax, and classification layer with output dimension. Early layers of the pretrained networks can be frozen (optionally) to increase training speed. For the CNN_new is new CNN network developed from scratch - trained with the dataset.

Institutions

Categories

Disease, X-Ray Imaging, Image Classification, Convolutional Neural Network, Deep Learning, Transfer Learning

Licence