Estimating the volume of biological structures from a single 2D image: considering apparent cross-sectional area as an alternative to the ellipsoid method

The volume of a biological structure may be estimated from photographs using its apparent cross-sectional area (ACA), or by measuring its dimensions from orthogonal viewpoints and treating it as an ellipsoid. With ACA half as many photographs are required to achieve results comparable to the ellipsoid method (EM); potentially simplifying data collection by reducing specimen handling time. ACA also facilitates the collection of additional morphological data for analyzing the shape of specimens using geometric morphometrics. To characterize ACA’s performance relative to EM, we carried out in-silico validations of ACA versus EM using virtual objects of known volume; first using a diversified set of 21 common objects, and subsequently, a set of 9 echinoderms scanned from wild-type specimens. The volume of each object was estimated twice, using pairs of images with orthogonal viewpoints. The relative and absolute performance (percent error) of both methods were evaluated. ACA performed comparably to EM across a broad range of object shapes and sizes. Both methods were sensitive to objects with low solidity and performed equally well for approximate spheres with high solidity. ACA tended to work as well or better than EM for objects with low circularity. Notably, in most cases, averaged ACA values also produced satisfactory or superior results relative to EM. Our results suggest that ACA represents a viable alternative to EM for estimating the volume of biological structures, especially those which deviate substantially from an ellipsoid form. Awareness of ACA’s applications and limitations, coupled with its ease of implementation, may encourage the inclusion of morphological metrics in experimental designs.