International Journal of Innovative Research in Computer and Communication Engineering
ISSN Approved Journal | Impact factor: 8.771 | ESTD: 2013 | Follows UGC CARE Journal Norms and Guidelines
| Monthly, Peer-Reviewed, Refereed, Scholarly, Multidisciplinary and Open Access Journal | High Impact Factor 8.771 (Calculated by Google Scholar and Semantic Scholar | AI-Powered Research Tool | Indexing in all Major Database & Metadata, Citation Generator | Digital Object Identifier (DOI) |
| TITLE | Hybrid Attention-Augmented CNN with GAN-Based Data Balancing for Robust Pneumonia Detection from Chest X-Ray Images |
|---|---|
| ABSTRACT | Pneumonia is a top cause of child deaths, with over 700,000 dying every year. It mainly affects kids in low- and middle-income countries where they struggle with lack of trained radiologists and diagnosis delays. Even though chest X-rays are the go-to diagnostic tools, they require a level of expertise that's not usually available at the point of care. We introduce HAA-CNN, a deep learning model designed to fix this problem. Three long-standing issues made us build our system the way it is: very uneven training data, neural network models that perform well but are opaque to clinicians, and compute-intensive architectures that can't be deployed in less developed regions. To get around the class imbalance, we trained a Spectral Normalization GAN (SGAN) to create synthetic normal chest X-rays that brought the training distribution back to an equal footing. For classification, we used a small depthwise separable CNN and a dual-branch spatial-channel attention mechanism to help the network focus on the most relevant regions of the image. Using Gradient-weighted Class Activation Mapping (Grad-CAM) we then visualized heatmaps that highlight the areas responsible for each classification, providing clinical interpretability. When tested against the Kermany pediatric X-ray benchmark, which consists of 5,863 images, HAA-CNN achieved 98.24% accuracy, 98.61% recall, 97.89% precision, a F1-score of 98.25%, and a AUC of 0.994. This achieved with an approximate parameter count of 3.8 million. We believe this combination of high accuracy, efficiency, and clinical interpretability offers a significant advance over the six previous systems that we compared against. |
| AUTHOR | PROF. USHA K, CHAITHRA M, CHANDANA M S, DIVYA MAJIGOUDRA, DIVYA S, GANASHREE M S Assistant Professor, Dept. of CSE, Jain Institute of Technology, Davangere, Karnataka, India UG Students, Dept. of CSE, Jain Institute of Technology, Davangere, Karnataka, India |
| VOLUME | 184 |
| DOI | DOI: 10.15680/IJIRCCE.2026.1405009 |
| pdf/9_Hybrid Attention-Augmented CNN with GAN-Based Data Balancing for Robust Pneumonia Detection from Chest X-Ray Images.pdf | |
| KEYWORDS | |
| References | [1] World Health Organization, "Pneumonia," WHO Fact Sheet, 2022. Available: https://www.who.int/news-room/fact-sheets/detail/pneumonia [2] Centers for Disease Control and Prevention, "Pneumonia," 2021. Available: https://www.cdc.gov/pneumonia [3] M. I. Neuman et al., "Variability in the interpretation of chest radiographs for the diagnosis of pneumonia in children," Journal of Hospital Medicine, vol. 7, no. 4, pp. 294-298, 2012. Https://doi.org/10.1002/jhm.955 [4] P. Rajpurkar et al., "CheXNet: Radiologist-level pneumonia detection on chest X-rays with deep learning," arXiv:1711.05225, 2017. [5] H. Almohab, "Deep Learning CNN for Pneumonia Detection: Advancing Digital Health in Society 5.0," Jurnal Ilmiah Profesi Pendidikan, vol. 10, no. 4, pp. 3787-3793, 2025. Https://doi.org/10.29303/jipp.v10i4.4001 [6] S. K. Anumula et al., "Deep Learning-Based CNN Model for Automated Detection of Pneumonia from Chest X-Ray Images," unpublished manuscript, 2025. [7] P. K. Dutta et al., "Deep Learning-Based Pneumonia Detection from Chest X-ray Images: A CNN Approach with Performance Analysis and Clinical Implications," arXiv:2510.00035, 2025. Https://doi.org/10.48550/arXiv.2510.00035 [8] S. R. Hole et al., "Chest X-ray Pneumonia Detection using Deep Learning," in Proc. ICBMESH, 2025. Https://doi.org/10.1109/ICBMESH66209.2025.11182215 [9] A. Roy Choudhury, "Pediatric Pneumonia Detection from Chest X-Rays: A Comparative Study of Transfer Learning and Custom CNNs," arXiv:2601.00837, 2025. Https://doi.org/10.48550/arXiv.2601.00837 [10] H. Slimi et al., "Trustworthy pneumonia detection in chest X-ray imaging through attention-guided deep learning," Scientific Reports, vol. 15, p. 40029, 2025. Https://doi.org/10.1038/s41598-025-23664-x [11] Z. Xu and H. Zhang, "Pneumonia detection from enhanced chest X-Ray images based on Double SGAN model," Scientific Reports, vol. 16, p. 9922, 2026. Https://doi.org/10.1038/s41598-026-39785-w [12] D. S. Kermany et al., "Identifying medical diagnoses and treatable diseases by image-based deep learning," Cell, vol. 172, no. 5, pp. 1122-1131, 2018. Https://doi.org/10.1016/j.cell.2018.02.010 [13] K. He et al., "Deep residual learning for image recognition," in Proc. IEEE CVPR, pp. 770-778, 2016. [14] G. Huang et al., "Densely connected convolutional networks," in Proc. IEEE CVPR, pp. 4700-4708, 2017. [15] M. Tan and Q. Le, "EfficientNet: Rethinking model scaling for CNNs," in ICML, pp. 6105-6114, 2019. [16] R. R. Selvaraju et al., "Grad-CAM: Visual explanations from deep networks via gradient-based localisation," in Proc. IEEE ICCV, pp. 618-626, 2017. [17] T. Miyato et al., "Spectral normalisation for generative adversarial networks," arXiv:1802.05957, 2018. [18] N. V. Chawla et al., "SMOTE: Synthetic minority over-sampling technique," JAIR, vol. 16, pp. 321-357, 2002. [19] K. Yu et al., "Artificial intelligence in healthcare," Nature Biomedical Engineering, vol. 2, no. 10, pp. 719-731, 2018. Https://doi.org/10.1038/s41551-018-0305-z [20] I. Goodfellow et al., "Generative adversarial networks," Communications of the ACM, vol. 63, pp. 139-144, 2020. |
