We are going to use kaggle image dataset (https://www.kaggle.com/datasets/trainingdatapro/computed-tomography-ct-of-the-chest) to illustrate the data processing and model preparation. This model will then be used for analysis & classification of CT Chest images.
We have 2 different functions included with this code sample where we will be using Transfer Learning as defined earlier with InceptionV3 model and the other one is a simple multi-layered CNN-DNN model.
The images in the dataset were trained using our simple multi-layered CNN-DNN model where with the 1st training of 15 epochs we saw an accuracy of 92.25% with validation accuracy fluctuating to be around 82.7%.
Pls feel free to modity the model architecture or tune the parameters to improve the training and validation accuracy.