Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Innovative Silicosis and Pneumonia Classification: Leveraging Graph Transformer Post-hoc Modeling and Ensemble Techniques

Bao Q. Bui, Tien T. T. Nguyen, Duy M. Le, Cong Tran, Cuong Pham

TL;DR

This work tackles three-class lung inflammation classification arising from silica exposure, bacteria, and viruses by introducing the SVBCX dataset and a Graph Transformer Post-Hoc (GTP) framework with Balanced Cross Entropy. The method combines a DNN encoder with a GTN to produce discriminative embeddings, and leverages ensemble learning to boost robustness, achieving a macro-F1 of 0.9749 and AUC-ROC above 0.99 across classes. Key contributions include a new CXR dataset tailored to occupational lung diseases, the novel GTP plugin architecture, and thorough evaluation including radiologist comparisons. The results demonstrate strong performance and practical potential for allied healthcare settings, especially for workers exposed to silica dust, with future work focusing on clinical integration and further methodological refinements.

Abstract

This paper presents a comprehensive study on the classification and detection of Silicosis-related lung inflammation. Our main contributions include 1) the creation of a newly curated chest X-ray (CXR) image dataset named SVBCX that is tailored to the nuances of lung inflammation caused by distinct agents, providing a valuable resource for silicosis and pneumonia research community; and 2) we propose a novel deep-learning architecture that integrates graph transformer networks alongside a traditional deep neural network module for the effective classification of silicosis and pneumonia. Additionally, we employ the Balanced Cross-Entropy (BalCE) as a loss function to ensure more uniform learning across different classes, enhancing the model's ability to discern subtle differences in lung conditions. The proposed model architecture and loss function selection aim to improve the accuracy and reliability of inflammation detection, particularly in the context of Silicosis. Furthermore, our research explores the efficacy of an ensemble approach that combines the strengths of diverse model architectures. Experimental results on the constructed dataset demonstrate promising outcomes, showcasing substantial enhancements compared to baseline models. The ensemble of models achieves a macro-F1 score of 0.9749 and AUC ROC scores exceeding 0.99 for each class, underscoring the effectiveness of our approach in accurate and robust lung inflammation classification.

Innovative Silicosis and Pneumonia Classification: Leveraging Graph Transformer Post-hoc Modeling and Ensemble Techniques

TL;DR

This work tackles three-class lung inflammation classification arising from silica exposure, bacteria, and viruses by introducing the SVBCX dataset and a Graph Transformer Post-Hoc (GTP) framework with Balanced Cross Entropy. The method combines a DNN encoder with a GTN to produce discriminative embeddings, and leverages ensemble learning to boost robustness, achieving a macro-F1 of 0.9749 and AUC-ROC above 0.99 across classes. Key contributions include a new CXR dataset tailored to occupational lung diseases, the novel GTP plugin architecture, and thorough evaluation including radiologist comparisons. The results demonstrate strong performance and practical potential for allied healthcare settings, especially for workers exposed to silica dust, with future work focusing on clinical integration and further methodological refinements.

Abstract

This paper presents a comprehensive study on the classification and detection of Silicosis-related lung inflammation. Our main contributions include 1) the creation of a newly curated chest X-ray (CXR) image dataset named SVBCX that is tailored to the nuances of lung inflammation caused by distinct agents, providing a valuable resource for silicosis and pneumonia research community; and 2) we propose a novel deep-learning architecture that integrates graph transformer networks alongside a traditional deep neural network module for the effective classification of silicosis and pneumonia. Additionally, we employ the Balanced Cross-Entropy (BalCE) as a loss function to ensure more uniform learning across different classes, enhancing the model's ability to discern subtle differences in lung conditions. The proposed model architecture and loss function selection aim to improve the accuracy and reliability of inflammation detection, particularly in the context of Silicosis. Furthermore, our research explores the efficacy of an ensemble approach that combines the strengths of diverse model architectures. Experimental results on the constructed dataset demonstrate promising outcomes, showcasing substantial enhancements compared to baseline models. The ensemble of models achieves a macro-F1 score of 0.9749 and AUC ROC scores exceeding 0.99 for each class, underscoring the effectiveness of our approach in accurate and robust lung inflammation classification.
Paper Structure (28 sections, 9 equations, 5 figures, 5 tables)

This paper contains 28 sections, 9 equations, 5 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Background and Motivation
  3. Organization and Notations
  4. Related work
  5. Radiologist-Guided Methods
  6. Traditional machine learning methods
  7. Deep learning based methods
  8. Benchmark Dataset
  9. Data collection
  10. Data processing
  11. Proposed method
  12. Problem formulation
  13. Graph transformer Post-Hoc design
  14. Loss functions
  15. Ensemble of models
  16. ...and 13 more sections

Figures (5)

  • Figure 1: Statistics of individuals diagnosed with silicosis pneumonia.
  • Figure 2: Sample images corresponding to four classification categories.
  • Figure 3: An overview of Graph Transformer Post-hoc architecture design.
  • Figure 4: The confusion matrices for the 4 classification classes correspond to 6 different ensemble cases.
  • Figure 5: The visualization of the attention maps corresponds to each image in the 4 classes and the chart represents the prediction confidence score of the DenseNet201-GTP model.