Machine Learning and Multiscale Methods in the Identification of Bivalve Larvae

This paper describes a novel application of support vec-tor machines and multiscale texture and color invariants to a problem in biological oceanography: the identifica-tion of 6 species of bivalve larvae. Our data consists of polarized color images of scallop and other bivalve larvae (between 2 and 17 days old) collected from the ocean by a shipboard optical imaging system of our design. Lar-vae of scallops, clams, and oysters are small (100 microns) with few distinguishing features when observed under stan-dard light microscopy. However, the use of polarized light with a full wave retardation plate produces a vivid color, bi-refringence pattern. The patterns display very subtle dif-ferences between species, often not discernable to human observers. We show that a soft-margin support vector ma-chine with Gaussian RBF kernel is a good discriminator on a feature set extracted from Gabor wavelet transforms and color distribution angles of each image. By constraining the Gabor center frequencies to be low, the resulting system can attain classification accuracy in excess of 90 % for ver-tically oriented images, and in excess of 80 % for randomly oriented images.