Bayesian convolutional neural network estimation for pediatric pneumonia detection and diagnosis
Vandecia Fernandes1, Geraldo Braz Junior1, Anselmo Cardoso de Paiva1
1Federal University of Maranhão, Applied Computing Group - NCA, Av. dos Portugueses, 1996, Campus do Bacanga, São Luís, Maranhão 65080-805, Brazil.
Insights
This study developed a convolutional neural network (CNN) using Bayesian optimization for accurate pneumonia detection and classification. The AI model achieved high accuracy, aiding in faster diagnosis for children.
Area of Science:
- Medical imaging analysis
- Artificial intelligence in healthcare
- Pediatric respiratory diseases
Background:
- Pneumonia is the leading cause of death in children under five globally.
- Accurate and rapid diagnosis is crucial for reducing pneumonia mortality.
- Computational methods can assist specialists and improve diagnostic accuracy.
Purpose of the Study:
- To develop an automated method for pneumonia detection and classification.
- To utilize convolutional neural networks (CNNs) for improved diagnostic speed and accuracy.
- To explore Bayesian optimization for efficient CNN architecture design.
Main Methods:
- A specific CNN architecture was constructed using Bayesian optimization.
- The method leveraged pre-trained networks for pneumonia detection.
- Classification of viral and bacterial pneumonia types was performed.
Main Results:
- The proposed CNN achieved 0.964 accuracy for pneumonia detection.
- Pneumonia type classification accuracy reached 0.957.
- The model demonstrated high performance without traditional image preprocessing.
Conclusions:
- Bayesian optimization is efficient for designing CNNs for pneumonia diagnosis.
- The proposed CNN architecture is effective for detecting and classifying pneumonia.
- The method's ability to perform without preprocessing highlights its efficiency and high performance.
Background And Objectives:
Pneumonia is a disease that affects the lungs, making breathing difficult. Nowadays, pneumonia is the disease that kills the most children under the age of five in the world, and if no action is taken, pneumonia is estimated to kill 11 million children by the year 2030. Knowing that rapid and accurate diagnosis of pneumonia is a significant factor in reducing mortality, acceleration, or automation of the diagnostic process is highly desirable. The use of computational methods can decrease specialists' workload and even offer a second opinion, increasing the number of accurate diagnostics.
Methods:
This work proposes a method for constructing a specific convolutional neural network architecture to detect pneumonia and classify viral and bacterial types using Bayesian optimization from pre-trained networks.
Results:
The results obtained are promising, in the order of 0.964 accuracy for pneumonia detection and 0.957 accuracy for pneumonia type classification.
Conclusion:
This research demonstrated the efficiency of CNN architecture estimation for detecting and diagnosing pneumonia using Bayesian optimization. The proposed network proved to have promising results, despite not using common preprocessing techniques such as histogram equalization and lung segmentation. This fact shows that the proposed method provides efficient and high-performance neural networks since image preprocessing is unnecessary.