PlexusNet: A neural network architectural concept for medical image classification

• Published in: Computers in biology and medicine Vol. 154; p. 106594
• Main Authors: Eminaga, Okyaz, Abbas, Mahmoud, Shen, Jeanne, Laurie, Mark, Brooks, James D., Liao, Joseph C., Rubin, Daniel L.
• Format: Journal Article
• Language: English
• Published: United States Elsevier Ltd 01.03.2023; Elsevier Limited
• Subjects: Accuracy; Adaptation, Physiological; Algorithms; Architecture; Artificial neural networks; Classification; Compact models; Computer vision; Convolutional neural networks; Datasets; Deep learning; Diagnostic Imaging; Diagnostics; Efficiency; Feature maps; Hypotheses; Image classification; Internal Medicine; Medical diagnosis; Medical imaging; Neural networks; Neural Networks, Computer; Other; PlexusNet; Source code; Standard scores; Deep learning; Computer vision; PlexusNet; Diagnostics; Compact models; Convolutional neural networks
• ISSN: 0010-4825; 1879-0534; 1879-0534
• DOI: 10.1016/j.compbiomed.2023.106594

State-of-the-art (SOTA) convolutional neural network models have been widely adapted in medical imaging and applied to address different clinical problems. However, the complexity and scale of such models may not be justified in medical imaging and subject to the available resource budget. Further increasing the number of representative feature maps for the classification task decreases the model explainability. The current data normalization practice is fixed prior to model development and discounting the specification of the data domain. Acknowledging these issues, the current work proposed a new scalable model family called PlexusNet; the block architecture and model scaling by the network's depth, width, and branch regulate PlexusNet's architecture. The efficient computation costs outlined the dimensions of PlexusNet scaling and design. PlexusNet includes a new learnable data normalization algorithm for better data generalization. We applied a simple yet effective neural architecture search to design PlexusNet tailored to five clinical classification problems that achieve a performance noninferior to the SOTA models ResNet-18 and EfficientNet B0/1. It also does so with lower parameter capacity and representative feature maps in ten-fold ranges than the smallest SOTA models with comparable performance. The visualization of representative features revealed distinguishable clusters associated with categories based on latent features generated by PlexusNet. The package and source code are at https://github.com/oeminaga/PlexusNet.git. [Display omitted] •Trade-off between model complexity and performance is poor in medical imaging.•Covariate shift and overparameterization are commonly neglected in medical imaging.•PlexusNet provides an optimized trade-off between model complexity and performance.•PlexusNet successfully mitigates covariate shift and overparameterization issues.•The image domain or modality can impact the optimal model architecture definition.