Pediatric Pneumonia Recognition Using an Improved DenseNet201 Model with Multi-Scale Convolutions and Mish Activation Function

Algorithms, Journal Year: 2025, Volume and Issue: 18(2), P. 98 - 98

Published: Feb. 10, 2025

Pediatric pneumonia remains a significant global health issue, particularly in low- and middle-income countries, where it contributes substantially to mortality children under five. This study introduces deep learning model for pediatric diagnosis from chest X-rays that surpasses the performance of state-of-the-art methods reported recent literature. Using DenseNet201 architecture with Mish activation function multi-scale convolutions, was trained on dataset 5856 X-ray images, achieving high performance: 0.9642 accuracy, 0.9580 precision, 0.9506 sensitivity, 0.9542 F1 score, 0.9507 specificity. These results demonstrate advancement diagnostic precision efficiency within this domain. By highest accuracy score compared other work using same dataset, our approach offers tangible improvement resource-constrained environments access specialists sophisticated equipment is limited. While need high-quality datasets adequate computational resources general consideration applications, model’s demonstrably superior establishes new benchmark delivery more timely precise diagnoses, potential significantly enhance patient outcomes.

Language: Английский

MSSFN: A multi-scale sequence fusion network for ct-based diagnosis of pulmonary complications

Neurocomputing, Journal Year: 2025, Volume and Issue: unknown, P. 129878 - 129878

Published: March 1, 2025

Language: Английский

Studying the Behavior of a Modified Deep Learning Model for Disease Detection Through X-ray Chest Images

International Journal of Advanced Computer Science and Applications, Journal Year: 2024, Volume and Issue: 15(5)

Published: Jan. 1, 2024

In modern medical diagnostics, Deep Learning models are commonly used for illness diagnosis, especially over X-ray chest images. approaches provide unmatched promise early identification, prognosis, and treatment evaluation across a range of illnesses, by combining sophisticated algorithms with large datasets. It is crucial to research these lead improved ones progress toward disease identification's precision, effectiveness, scalability. This paper presents the study CNN+VGG19 architecture (subsets machine learning), both before after its modification. The same dataset existing modified compare metrics under conditions. They compared using like loss, accuracy, sensitivity, AUC. These display lower values in updated model than original one. numbers demonstrate occurrence overfitting phenomenon, which most likely result model's increased complexity small dataset. noise images included may also be cause. As result, it can stated that regularization techniques should applied; otherwise, layers extraction classification not added prevent overfitting.

Language: Английский