ARCHIVES
VOL. 11, ISSUE 1 (2026)
Development of an Xception-based convolutional autoencoder (Xnetcae) feature extraction technique for enhanced lung cancer detection
Authors
Olaleye O J, Olabiyisi S O, Ismaila W O, Achas M J, Ashade B T
Abstract
This study presents an enhanced deep learning
feature extraction technique for early detection of lung cancer from low-dose
computed tomography (LDCT) images. The Xception network was integrated into a
standard convolutional autoencoder (CAE) architecture and the resultant
Xception-based Convolutional Autoencoder (XnetCAE) was trained, validated and
tested with a portion (20,000) of lung scan images from the Lung Image Database
Consortium-Image Database Resource Initiative (LIDC-IDRI) dataset. The performance
of the XnetCAE was evaluated and compared with that of a standard CAE using
reconstruction loss, accuracy, precision, recall, F1-score and computation time
metrics. With the XnetCAE technique, the reconstruction loss, accuracy,
precision, recall, F1-score and computation time are 0.16, 71.97%, 0.79, 0.72,
0.75 in 36 seconds, while the standard CAE achieved 0.30, 65%, 0.65, 0.65, 0.65
in 48 seconds, respectively. These results show that the XnetCAE offers a
superior performance over standard CAE.
Download
Pages:7-18
How to cite this article:
Olaleye O J, Olabiyisi S O, Ismaila W O, Achas M J, Ashade B T "Development of an Xception-based convolutional autoencoder (Xnetcae) feature extraction technique for enhanced lung cancer detection". International Journal of Advanced Science and Research, Vol 11, Issue 1, 2026, Pages 7-18
Download Author Certificate
Please enter the email address corresponding to this article submission to download your certificate.