Jul 2026· International Conference on Information and Communicatiaon Technology· pp. 1-6· 0 citations· 30 references
Abstract
Attention mechanisms are widely adopted in convolutional neural networks to improve feature representation, yet their consistent effectiveness in medical image classification remains uncertain. This study systematically evaluates channel and spatial attention mechanisms integrated into a ResNet50 architecture for chest X-ray classification tasks. Three attention configurations are examined, namely Squeeze-and-Excitation (SE), Convolutional Block Attention Module (CBAM), and a hybrid combination of both, using three publicly available chest X-ray datasets: Pneumonia Chest X-ray, COVID-QU-Ex, and Tuberculosis Chest X-ray. All model variants are trained under identical preprocessing and optimization settings to ensure a controlled and fair comparison. Each experiment is repeated across multiple random seeds to reduce the influence of stochastic variability, and final performance is reported as the mean across runs. Model effectiveness is assessed using accuracy, precision, recall, and F1 score, followed by paired t-tests to determine the statistical significance of observed differences. Quantitative results indicate that attention-augmented models produce marginal performance changes relative to the ResNet50 baseline across all three datasets. Statistical testing confirms that none of the observed improvements reach significance, with all p-values exceeding 0.05. These findings suggest that attention mechanisms do not function as universally effective performance boosters in chest X-ray classification. Instead, their role appears to be supportive in nature, contributing to feature representation stability rather than driving substantial predictive gains. Dataset characteristics and attention placement are identified as key factors influencing the degree of benefit observed.
Pneumonia remains one of the leading causes of morbidity and mortality worldwide, particularly among children, older adults, and immunocompromised individuals. Although chest X-ray (CXR) imaging is widely used for pneumonia diagnosis, manual interpretation is time-consuming, subjective, and highly dependent on radiologist expertise. Deep learning has shown promising performance for automated pneumonia classification; however, class imbalance remains a major challenge that can lead to biased predictions and reduced model generalization. Therefore, this study investigates the effectiveness of image augmentation and class weighting for handling class imbalance in pneumonia classification using chest X-ray images. The main contribution of this study is a systematic comparison of four experimental scenarios: Baseline, Augmentation Only, Class Weighting Only, and Hybrid (Image Augmentation and Class Weighting) implemented on a ResNet50 architecture integrated with the Convolutional Block Attention Module (CBAM). Experiments were conducted using the publicly available Chest X-Ray Images (Pneumonia) dataset from Kaggle, comprising 1,583 Normal and 4,273 Pneumonia images. Model performance was evaluated using Accuracy, ROC-AUC, Precision, Recall, F1-Score, confusion matrix analysis, and Youden’s J-Statistic for threshold optimization. The Hybrid model achieved the best overall performance, with an Accuracy of 95.74%, a ROC-AUC of 98.83%, a Macro Precision of 96.01%, a Macro Recall of 93.13%, and a Macro F1-Score of 94.44%. Moreover, the number of false negative predictions decreased from 25 in the Baseline model to 5 in the Hybrid model. These findings demonstrate that integrating image augmentation and class weighting within the ResNet50-CBAM framework effectively mitigates class imbalance and improves the reliability of automated pneumonia classification.
Kafilah Akhmad Fatahillah, T. H. Saragih, D. Kartini et al.· Indonesian Journal of Electr...· 0 citations
This study develops a lightweight Convolutional Neural Network (CNN) model for pneumothorax
detection in chest X-rays, targeting resource-limited healthcare environments. The proposed
architecture integrates EfficientNetB2 for feature extraction, Long Short-Term Memory (LSTM)
layers for sequential spatial modeling, and a Multi-Head Attention mechanism to enhance focus
on critical regions. Trained on the SIIM-ACR Pneumothorax dataset (2,027 images), the model
employs data augmentation, balancing, and preprocessing to address class imbalance and
variability. The evaluation results demonstrate strong performance, achieving 86% accuracy,
93% recall, and a 0.91 AUC-ROC score, outperforming baseline models like ResNet-50 and
MobileNet-v2. The model’s clinical applicability is further validated through Gradient-weighted
Class Activation Mapping (Grad-CAM) visualizations, highlighting lesion-specific regions, and a
user-friendly GUI for real-world deployment. By optimizing computational efficiency while
maintaining diagnostic accuracy, this work bridges the gap between deep learning and practical
medical applications, particularly in underserved regions. Ethical, legal, and environmental
considerations, including GDPR compliance and energy-efficient design, are systematically
addressed to ensure responsible AI deployment.
Miracle Ugomma Anunobi· International Journal of Med...· 0 citations
A hybrid ensemble learning approach to classify chest X-ray images into four classes—Normal, COVID-19, Pneumonia, and Tuberculosis exhibited high sensitivity in detecting Tuberculosis with considerable stability in classifying Normal, COVID-19, and Pneumonia.
Abdul Rehman Khan Tareen, Muhammad Laiq Ur Rahman Shahid, Muhammad Hamza Zafar et al.· Allied Medical Research Jour...· 0 citations
Lung diseases such as pneumonia, tuberculosis, and the current coronavirus (COVID-19) are significant causes of morbidity and mortality in the world. Early diagnosis and accurate diagnosis based on chest X-rays (CXR) is very important for effective treatment, but manual interpretation takes time and can be prone to errors. It is well known that the conventional deep learning(DL) models are incomplete to capture not only the finegrained but also the global pathological features of the heterogeneous CXRs, which lead to a suboptimal classification performance. This paper proposes to build a robust automated framework for classification of lung diseases which enables multiscale feature extraction along with attention mechanism interpreted for lung disease. This study proposes a multi-scale DenseNetwork of a hybrid shuffle-spatial attention (HSSA) module, in order to suppress the non-important features of medical images and capture discriminative features at multiple resolutions and emphasize on clinically important regions. The model was trained and tested against the NIH ChestX-ray14 dataset using some preprocessing, augmentation, and end-to-end supervised learning. Experimental results show that it achieves better performance compared to 5 state-of-the-art models with 92.8% accuracy, 91.1% F1-score and 95.4% AUC-ROC with attention maps and visual interpretability. The proposed framework of MS-DenseNet + HSSA shows significant improvement in terms of automatic lung disease detection from CXRs, it provides a reliable, explainable, and clinically applicable method for the purpose of analysing CXRs for the radiologist community and also willing to be incorporated in real-world diagnostic processes.
V. Nandhini, R. Parameswari· International Conference on...· 0 citations
Background: Deep learning models, particularly convolutional neural networks (CNNs), have shown promising performance for pneumonia detection using chest X-ray images. However, the impact of preprocessing, architecture selection, data augmentation, and ensemble strategies has not been systematically evaluated. This study investigated how these factors affect model robustness and diagnostic performance. Methods: A public pediatric chest X-ray dataset was used to systematically evaluate pixel normalization methods, six CNN architectures, progressive data augmentation strategies for class imbalance, and both feature-level and decision-level ensemble approaches. Model performance was assessed by considering not only overall classification accuracy but also clinically relevant risk metrics, particularly false-negative rates. Results: Pixel normalization to the 0–1 range improved model convergence, while Xception and InceptionV3 achieved the best overall performance. Model-specific augmentation strategies were more effective than a fixed 1:1 class ratio for reducing false negatives. Feature-level ensembles tended to overfit, whereas decision-level ensembles provided more stable but only modest performance improvements. Conclusions: These findings demonstrate that reliable medical AI systems require systematic optimization of preprocessing techniques, model architecture, data augmentation strategies, and clinically meaningful evaluation metrics rather than maximizing a single performance indicator. The proposed framework provides practical guidelines for developing robust deep learning models for pneumonia diagnosis in clinical settings.
YongJun Kim, Ji-Yeoun Lee· BioMedInformatics· 0 citations
In order to diagnose respiratory disorders such as pneumonia and COVID-19, X-ray imaging of the chest is essential. On the other hand, radiologists could differ in their approaches and the amount of time it takes to manually analyze radiographic pictures. Recent innovations in deep learning have greatly enhanced automated examination of medical pictures, although conventional convolutional neural networks often miss long-range spatial interactions in complicated pulmonary patterns. To overcome this drawback, this work suggests a Vision Transformer (ViT)-based feature extraction system embedded in the GenMAT-Net framework to classify X-ray images of the chest automatically. The proposed method initially achieves lung region segmentation to isolate clinically significant regions of the image. After that, the pictures are divided they are converted into patch embeddings after being converted into fixed-size patches. To keep the spatial relationships, they are augmented using positional encoding. To acquire global contextual information and multifarious feature linkages throughout the lung regions, these embeddings are trained using numerous layers of transformer encoders that consist of multi-head self-attention and feed forward networks. The semantic features representations at high-level are then optimized and used to classify the disease into three categories, namely, normal, pneumonia, and COVID-19. As shown by experimental assessment, feature extraction based on transformers offers better contextual representation of the lung abnormalities, which allows the effective identification of the pathological patterns, i.e., ground-glass opacities, consolidations, and diffuse infiltrates. The proposed framework also embraces interpretability based on attention mechanisms, where visualization of regions leading to the diagnostic decision can be done. In general, the technique for extracting features based on the application of the Vision Transformer can increase diagnostic reliability and provide a promising solution to intelligent computer-aided diagnostic systems in the medical imaging field.
P. V. Naga Lakshmi, K. Vedavathi· ITM Web of Conferences· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
