A High Throughput and User-Friendly Computational Tool for Quantifying and Visualizing Symmetry Variation in Image Datasets
by Liam D. O’Brien, Punit Gandhi, Maria-Veronica Ciocanel & Adriana T. Dawes
Read the paper
Symmetry is a fundamental property of living systems. Researchers typically use discrete classifications such as radial and bilateral symmetry for biological images, but these categories ignore the fact that biological symmetries are seldom exact. We present an automated computational pipeline for quantifying and visualizing symmetry in image datasets. This method can capture the full spectrum of symmetry variation in living systems, can minimize user bias, and can be applied to images with diverse symmetry types. We illustrate its versatility by applying it to real flower images and to simulated ferns. In the floral data, our analysis shows that distantly-related flowers exhibit similar symmetries, suggesting convergent evolution.
Image Description:
