Lung cancer detection in CT scans using extracted nodule features and precise multistage segmentation with deep learning based ensemble model

Deep Learning-Based Lung Cancer Diagnosis and Classification Using CT Images

One of the main causes of cancer-related fatalities globally is lung cancer, and increasing patient survival requires early diagnosis. Lung cancer screening frequently uses computed tomography (CT) imaging, although manual interpretation can be laborious and reliant on radiologist skill. A deep learning-based framework for CT image-based lung cancer diagnosis and classification is presented in this work. Lung CT scans are divided into three groups by the suggested system: benign, malignant, and normal. Preprocessing methods including Intensity-based feature enhancement, Contrast Limited Adaptive Histogram Equalization (CLAHE), and normalizing are used to improve image quality and highlight significant features. Both local texture features and global contextual information are extracted from CT images using a hybrid deep learning model that combines ResNet18 and Vision Transformer (ViT). Additionally, data augmentation is used to enhance model generalization and lessen class imbalance. The Hybrid-RViT model, which combined ResNet18 and ViT, outperformed the competition with 97.13% accuracy. These findings show that the suggested methodology can efficiently facilitate automated lung cancer diagnosis and help physicians make decisions more quickly and accurately.

Efficientnet-B0-Based Deep Learning Framework for Automated Lung Cancer Detection Using Chest Computed Tomography Images

However, early-stage lung cancer is one of the major causes of death from cancer due to its symptoms being absent or hard to detect by conventional clinical analysis. The late diagnosis of this disease causes the effectiveness of the treatment to become poor and low survival rates. Traditional interpretation of medical imaging involves manual analysis which makes it burdensome, prone to variation between observers, and inefficient for screening in cases where many chest CT images need to be analyzed. This calls for efficient computer-based intelligent diagnostics systems that can provide effective, reliable, and fast classification results. In this paper, an EfficientNet-B0-based deep learning model for automated lung cancer classification using chest CT images is proposed. Preprocessing steps such as normalization, lung segmentation, resizing, and data augmentation are used. The custom dataset, specifically PulmoVision Lung Cancer CT Dataset (PLCD-2026), contains balanced image samples that correspond to healthy lungs, benign nodules, and malignancies in order to provide sufficient conditions for effective model training. EfficientNet-B0 learns discriminative anatomical and pathological features and remains computationally efficient by means of optimized network scaling. High classification capabilities of the developed approach can be proved by means of the evaluation procedure conducted in terms of accuracy, precision, recall, and F1-score, which proves reliable identification of different pulmonary diseases. Grad-based visualization is additional confirmation of proper localization of suspicious areas, contributing to the interpretability of the classification process. The developed framework provides an efficient tool for lung cancer detection, helps doctors to make decisions, saves their time, and ensures consistent clinical assessment.

Deep learning driven early detection of lung cancer from CT images using transfer learning approaches

The proposed system demonstrates potential as an assistive tool for automated lung cancer screening, warranting further validation on larger and multi-institutional datasets before clinical application.

Deep Learning Classification Based Fast Segmentation for Detection of Lung Cancer

Experimental results demonstrate that the proposed DLC-FS framework significantly outperforms existing state-of-the-art methods in both detection accuracy and computational efficiency, making the system suitable for real-time clinical applications.

Multi-class lung cancer detection and classification from CT scans using CNN transformer

LungCNET: A High-Performance Deep CNN Model for Lung Cancer Detection Evaluated Against Widely Used CNN Benchmarks

Lung cancer arises from mutations in lung cells, disrupting their normal growth cycle and leading to uncontrolled cell division. These rapidly dividing cells lose function and fail to form healthy lung tissue. Several factors contribute to the difficulty of diagnosing and classifying lung nodules, including the high degree of morphological heterogeneity and overlapping characteristics between benign and malignant nodules. Recently, deep learning models have been used in computed tomography (CT)-based lung nodule diagnosis and have demonstrated diagnostic efficiency comparable to that of radiologists. This study introduces LungCNET, a high-performance multi-layer deep convolutional neural network trained on chest CT images to improve lung lesion classification efficiency and accuracy significantly. The data for the Lung Cancer convolutional neural network (LungCNET) were derived from the IQ-OTH/NCCD CT scan dataset (1097 images from 110 cases), split into training (767 images), validation (109) and a held-out test partition (221) that played no role in training or model selection. This dataset encompasses three diagnostic categories: benign, malignant, and normal lung tissues. LungCNET was evaluated against fine-tuned benchmark models that are both established and widely used, spanning architectures introduced between 2014 and 2024, including VGG16, ResNet50, InceptionV3, MobileNetV2, and YOLOv11. On the held-out test partition, LungCNET reached a macro-averaged F1-score of 95.19%, with VGG16 at 94.09% and InceptionV3 at 92.28%; these three models performed comparably, and the separation between them is small relative to the resolution of a test set of this size. LungCNET was, however, the only model to exceed 90% F1-score across all three diagnostic classes simultaneously, and recorded the highest F1-score on the benign class (91.0%), the smallest and most frequently misclassified category, where two of the six models failed entirely. These results support LungCNET as a candidate tool for lung cancer diagnosis, subject to validation on larger and independently sourced datasets.