How many specimens make a sufficient training set for automated three-dimensional feature extraction?

Abstract

Deep learning has emerged as a robust tool for automating feature extraction from three-dimensional images, offering an efficient alternative to labour-intensive and potentially biased manual image segmentation methods. However, there has been limited exploration into the optimal training set sizes, including assessing whether artficial expansion by data augmentation can achieve consistent results in less time and how consistent these benefits are across different types of traits. In this study, we manually segmented 50 planktonic foraminifera specimens from the genus Menardella to determine the minimum number of training images required to produce accurate volumetric and shape data from internal and external structures. The results reveal unsurprisingly that deep learning models improve with a larger number of training images with eight specimens being required to achieve 95% accuracy. Furthermore, data augmentation can enhance network accuracy by up to 8.0%. Notably, predicting both volumetric and shape measurements for the internal structure poses a greater challenge compared with the external structure, owing to low contrast differences between different materials and increased geometric complexity. These results provide novel insight into optimal training set sizes for precise image segmentation of diverse traits and highlight the potential of data augmentation for enhancing multivariate feature extraction from three-dimensional images.

Morphological simulation tests the limits on phenotype discovery in 3D image analysis

“…Augmentation often involves applying simple transforms such as changes in voxel intensity, cropping, rotations, and resizing ( e.g. , Hussain et al ., 2017; Nanni et al ., 2021; Mulqueeney et al ., 2024). Additionally, methods for representing more complex shape differences based on the training data have been implemented using transforms derived from PCA of the modes of shape change in a dataset ( e.g.…”

“…Data augmentation is an important method widely used to improve the accuracy and robustness of deep learning models by making changes to the data that are representative of the differences that are expected to be observed between subjects that are not in the training dataset (e.g., Hussain et al, 2017;Tustison et al, 2019;Nanni et al, 2021). Augmentation often involves applying simple transforms such as changes in voxel intensity, cropping, rotations, and resizing (e.g., Hussain et al, 2017;Nanni et al, 2021;Mulqueeney et al, 2024). Additionally, methods for representing more complex shape differences based on the training data have been implemented using transforms derived from PCA of the modes of shape change in a dataset (e.g., Nanni et al, 2021).…”