Histopathological Image based Classification of Lungs Cancer using Convolutional Recurrent Neural Networks

Early diagnosis of Lung cancer is critical and can improve patient prognosis, increasing the likelihood of successful treatment. This research primarily focuses on the identification of a better-performing diagnostic deep-learning model that can assist histopathologists in detecting the subtypes of lung cancer at an early stage. Histopathological imaging presents microscopic medical samples of surgical specimens, which are significant for the identification of the disease. The two convolutional recurrent neural networks proposed in this paper are the Convolutional Neural Network-Gated Recurrent Unit and the Convolutional Neural Network + Long Short-Term Memory. In order to capture crucial patterns for classifying lung cancer, convolutional recurrent neural networks use convolutional neural networks to perform features extraction using the histopathology images. These characteristics are further utilized by recurrent neural networks. This research has compared models, namely - Convolutional Neural Network, Convolutional Neural Network-Gated Recurrent Unit and Convolutional Neural Network + Long Short-Term Memory. The research identifies the Convolutional Neural Network + Long Short-Term Memory network as the best network, achieving an accuracy of 97.93%. Overall, the efficient classification represents an essential step forward in the early diagnosis of lung cancer.