Convolutional neural network-based PSO for lung nodule false positive reduction on CT images. (August 2018)

• Record Type: Journal Article
• Authors: da Silva, Giovanni Lucca França; Valente, Thales Levi Azevedo; Silva, Aristófanes Corrêa; de Paiva, Anselmo Cardoso; Gattass, Marcelo
• Is Part Of: Computer methods and programs in biomedicine. Volume 162(2018)
• Page Start: 109
• Page End: 118
• Publication Date: 2018-08
• Subjects: Medical images -- Lung nodules -- False positive reduction -- Deep learning -- Convolutional neural network -- Particle swarm optimization; Medicine -- Computer programs -- Periodicals; Biology -- Computer programs -- Periodicals; Computers -- Periodicals; Medicine -- Periodicals; Médecine -- Logiciels -- Périodiques; Biologie -- Logiciels -- Périodiques; Biology -- Computer programs; Medicine -- Computer programs; Periodicals; Electronic journals; 610.28
• DOI: 10.1016/j.cmpb.2018.05.006
• Languages: English
• ISSNs: 0169-2607

Abstract

Highlights: This paper proposes a methodology to reduce lung nodule false positive on computed tomography scans. The proposed methodology uses a convolutional neural network in conjunction with the particle swarm optimization algorithm. The best result obtained was 97.62% of accuracy, 92.20% of sensitivity, 98.64% of specificity and AUC of 0.955 in the LIDC-IDRI database. Abstract: Background and objective: Detection of lung nodules is critical in CAD systems; this is because of their similar contrast with other structures and low density, which result in the generation of numerous false positives (FPs). Therefore, this study proposes a methodology to reduce the FP number using a deep learning technique in conjunction with an evolutionary technique. Method: The particle swarm optimization (PSO) algorithm was used to optimize the network hyperparameters in the convolutional neural network (CNN) in order to enhance the network performance and eliminate the requirement of manual search. Results: The methodology was tested on computed tomography (CT) scans from the Lung Image Database Consortium and Image Database Resource Initiative (LIDC-IDRI) with the highest accuracy of 97.62%, sensitivity of 92.20%, specificity of 98.64%, and area under the receiver operating characteristic (ROC) curve of 0.955. Conclusion: The results demonstrate the high performance-potential of the PSO algorithm in the identification of optimal CNN hyperparameters for lung nodule candidate classification into nodules and non-nodules, increasing the sensitivity rates in the FP reduction step of CAD systems.