Lung Cancer Classification with Histopathological Images using Otsu’s Method and CNN

Since lung cancer is one of the leading reasons for cancer-associated fatalities worldwide, early and precise identification is crucial to enhancing patient outcomes. Early identification performs a significant character in the survival rate and treatments of lung cancer patients. This paper examines the efficacy of utilizing Otsu’s image segmentation technique in combination with a Convolutional Neural Network (CNN) to classify lung cancer histopathology pictures. This research analysis is carried out by using the publicly released benchmark lungs tissue histopathological imaging dataset LC25000. The efficacy of Otsu’s segmentation on CNN’s F1-score, precision, recall, and accuracy are among the metrics utilized to compare the suggested method’s productiveness. The findings show that using Otsu’s approach greatly improves the model’s productiveness; with segmentation, the architecture achieves an accuracy of $\mathbf{9 1 \%}$. This demonstrates that the CNN with Otsu’s gives a better approach for classifying or detecting lung cancer.