<![CDATA[Cancer diagnosis from histopathological images is a crucial but complex task that can be improved through artificial intelligence. This study aims to design and evaluate a deep learning model for the automatic classification of five classes of histopathological networks of lung and colon cancer. The methodology used is to train and compare two neural network architectures, namely a custom Convolutional Neural Network (CNN) and a custom Recurrent Neural Network (RNN), on a balanced public dataset consisting of 25,000 images. The dataset was divided into training data (80%), validation data (10%), and testing data (10%) to ensure objective evaluation. The experimental results showed that the CNN model was significantly superior, achieving an accuracy of 97.52% on the test data, compared to the RNN, which only achieved 95.12%. Further analysis of the CNN model revealed very high classification performance across most classes, with an average F1 score of 98%, although it was found to have some difficulty distinguishing between two morphologically similar subtypes of lung cancer. It is concluded that the specifically designed CNN architecture is a highly effective and reliable approach for histopathological image classification, with strong implications as a potential diagnostic tool to accelerate and improve accuracy in clinical pathology practice.]]>
Copyrights © 2025
