COMPARATIVE ANALYSIS OF CUSTOMIZED CNN AND TEACHABLE MACHINE FOR PNEUMONIA DETECTION IN CHEST X-RAY IMAGES

Parth R. Dave; Yogesh B. Patel; Shivangkumar R. Patel

doi:10.29121/shodhkosh.v5.i3.2024.5173

Authors

Parth R. Dave Assistant Professor, Computer Engineering Department, L. D. College of Engineering- Ahmedabad, Gujarat
Yogesh B. Patel Assistant Professor, Computer Engineering Department, Govt. Engineering College – Patan, Gujarat
Shivangkumar R. Patel Assistant Professor, Computer Engineering Department, Govt. Engineering College – Modasa, Gujarat

DOI:

https://doi.org/10.29121/shodhkosh.v5.i3.2024.5173

Abstract [English]

Pneumonia is a serious and potentially life-threatening lung infection that necessitates prompt and accurate diagnosis to ensure effective treatment and reduce the risk of complications or fatalities. Traditional diagnosis through chest X-ray interpretation can be challenging due to variability in radiologist expertise and the subtle nature of some pneumonia-related abnormalities. With the rise of artificial intelligence in healthcare, deep learning techniques—particularly Convolutional Neural Networks (CNNs)—have shown great promise in automating and enhancing diagnostic accuracy in medical imaging. This study conducts a comparative evaluation of a custom-built CNN model and Google's Teachable Machine, a user-friendly no-code machine learning platform, for binary classification of chest X-ray images into two categories: "Normal" and "Pneumonia." Both models were trained and tested on the same dataset to ensure a fair comparison. The experimental outcomes reveal that the scratch-built CNN model significantly outperforms the Teachable Machine model in key performance metrics including accuracy, recall, precision, and F1-score. These results highlight the superiority of tailored deep learning architectures designed specifically for the complexities of medical image analysis. The findings suggest that while generalized machine learning tools offer accessibility, they may fall short in critical diagnostic tasks where specialized models can better capture domain-specific features for improved clinical decision-making.

References

World Health Organization, “Pneumonia,” World Health Organization, 2023. [Online]. Available: https://www.who.int/news-room/fact-sheets/detail/pneumonia

M. L. Giger, “Machine Learning in Medical Imaging,” J. Am. Coll. Radiol., vol. 15, no. 3, pp. 512–520, 2018. DOI: https://doi.org/10.1016/j.jacr.2017.12.028

J. R. Zech, M. A. Badgeley, M. Liu, A. B. Costa, J. J. Titano, and E. K. Oermann, “Variable generalization performance of a deep learning model to detect pneumonia in chest radiographs: A cross-sectional study,” PLoS Med., vol. 15, no. 11, pp. 1–17, 2018. DOI: https://doi.org/10.1371/journal.pmed.1002683

G. Litjens et al., “A survey on deep learning in medical image analysis,” Med. Image Anal., vol. 42, pp. 60–88, 2017. DOI: https://doi.org/10.1016/j.media.2017.07.005

D. Kermany, K. Zhang, and M. Goldbaum, “Labeled Optical Coherence Tomography (OCT) and Chest X-Ray Images for Classification,” Mendeley Data, vol. 2, no. 2, 2018. [Online]. Available: https://data.mendeley.com/datasets/rscbjbr9sj/2

Google, “Teachable Machine,” [Online]. Available: https://teachablemachine.withgoogle.com/

N. Anwar et al., “Evaluation of Google's Teachable Machine in Medical Image Classification,” in Proc. IEEE Int. Conf. Health Informatics (ICHI), 2021, pp. 317–324.

P. Rajpurkar et al., “CheXNet: Radiologist-level pneumonia detection on chest X-rays with deep learning,” arXiv preprint arXiv:1711.05225, 2017.

D. S. Kermany et al., “Identifying Medical Diagnoses and Treatable Diseases by Image-Based Deep Learning,” Cell, vol. 172, no. 5, pp. 1122–1131.e9, 2018. DOI: https://doi.org/10.1016/j.cell.2018.02.010

K. Simonyan and A. Zisserman, “Very Deep Convolutional Networks for Large-Scale Image Recognition,” arXiv preprint arXiv:1409.1556, 2014.

K. He, X. Zhang, S. Ren, and J. Sun, “Deep Residual Learning for Image Recognition,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2016, pp. 770–778. DOI: https://doi.org/10.1109/CVPR.2016.90

M. Tan and Q. V. Le, “EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks,” in Proc. Int. Conf. Mach. Learn. (ICML), 2019, pp. 6105–6114.

H. Zhou et al., “Analysis of Over parameterization in Transfer Learning,” in Proc. Int. Conf. Learn. Represent. (ICLR), 2020.

J. Wu, Y. Zhao, and S. Du, “Edge computing-enabled mobile object tracking for smart surveillance system,” IEEE Internet Things J., vol. 6, no. 6, pp. 9866–9878, Dec. 2019.

R. Fan et al., “A Brief Review of Deep Learning Methods for Monocular Object Pose Detection and Tracking,” IEEE Trans. Intell. Transp. Syst., vol. 23, no. 9, pp. 17029–17044, 2022.

A. Sharma et al., “Robust Object Detection under Adverse Weather Conditions,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. Workshops (CVPRW), 2019, pp. 1164–1172.

S. Zhang et al., “Hybrid Features Based on CNN and HOG for Cervical Cell Classification,” IEEE Access, vol. 6, pp. 22491–22503, 2018.

M. Sandler et al., “MobileNetV2: Inverted Residuals and Linear Bottlenecks,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 2018, pp. 4510–4520. DOI: https://doi.org/10.1109/CVPR.2018.00474

Y. Zhang et al., “YOLO-Based Colorectal Cancer Cell Detection Using Histopathological Images,” IEEE Access, vol. 8, pp. 68133–68140, 2020.

Z. Guefrachi et al., “Diabetic Retinopathy Screening Using Deep Learning: A Review,” Comput. Biol. Med., vol. 136, 104754, 2021.

V. Nair and G. E. Hinton, "Rectified Linear Units Improve Restricted Boltzmann Machines," in Proc. 27th Int. Conf. Mach. Learn. (ICML), 2010, pp. 807–814.