A COMPACT XCEPTION–GCN FRAMEWORK FOR ACCURATE LUNG CANCER SUBTYPE CLASSIFICATION FROM HISTOPATHOLOGY IMAGES

Read full text Summarise with VIRAT
Authors: DIP ACHARYA, SUBRATA SINHA
Cited by: 0
References: 4

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Accurate classification of lung cancer subtypes from histopathological images is critical for clinical diagnosis, yet existing convolutional neural networks (CNNs) inadequately capture relational dependencies among samples.To address this limitation, this work proposes a compact hybrid framework that integrates a pretrained Xception network for deep feature extraction with a Graph Convolutional Network (GCN) for relational learning.Extracted features are standardized, reduced using Principal Component Analysis, and structured into a similarity-based graph for classification.The proposed model is evaluated on a lung histopathology dataset comprising adenocarcinoma, squamous cell carcinoma, and non-malignant tissues, achieving classification accuracy and ROC-AUC values close to 99%.The framework offers an efficient and accurate CNN-GCN solution suitable for practical digital pathology applications.

Fields of Study & Subjects

AI in cancer detection Radiomics and Machine Learning in Medical Imaging Advanced Neural Network Applications

Publication details

Journal / source
Science and Culture
Volume
92
Publication date
17 March 2026
Language
EN