Abstract :

Lung Cancer Is The Foremost Cause Of Cancer-related Mortality, Requiring Prompt And Precise Diagnosis To Enhance Patient Outcomes. Manual Interpretation Of Computed Tomography (CT) Scans And Traditional Deep Learning Techniques Frequently Inadequately Identify Multi-scale Features And Accurately Localize Lesions. Experiments Are Performed Using Two Publicly Accessible Datasets: The IQ-OTH/NCCD Lung Cancer Dataset And The Chest CT-Scan Images Dataset. The Suggested Attention-enhanced Hybrid CNN–ViT Framework Amalgamates ResNet50, DenseNet169, EfficientNetV2-Medium, ConvNeXt-Base, InceptionNeXt-Base, MobileViT-Small, ConViT-Base, Swin-Base, MaxViT-Base, And DeiT3-Base For Classification, In Conjunction With YOLOv5, YOLOv8, YOLOv9, And YOLOv11 For Detection. Preprocessing Encompasses Image Resizing To 299×299, Data Augmentation, Tensor Normalization, And Stratified Data Partitioning, Whereas YOLO Datasets Are Organized With Bounding Box Annotations. GradCAM Produces Heatmaps That Emphasize Significant Areas, While A Flask-based Interface Facilitates Comprehensive User Interaction. ConvNeXt-Base Attains The Maximum Classification Accuracy Of 99.09% On The IQOTH/NCCD Dataset, Whereas InceptionNeXt-Base Achieves 99.01% Accuracy On The Chest CT-scan Dataset. YOLOv5 Attains The Highest Mean Average Precision (mAP) Of 72.8% For Detection. The Approach Exhibits Enhanced Resilience, Equitable Performance, And Interpretable Predictions By The Integration Of Categorization And Detection Within A Cohesive System.

Published:

08-4-2026

Issue:

Vol. 26 No. 4 (2026)

Page Nos:

1653-1666

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite