Optimizing multiscale texture invariants for the identification of bivalve larvae

This paper describes a novel application of multiscale texture invariants and statistical learning theory to the identification of 6 species of bivalve larvae in biological oceanography. 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. Larvae of scallops, clams, and oysters are small (100 microns) with few distinguishing features when observed under standard 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 differences between species, often not discernable to human observers. Our texture invariants are extracted from Gabor wavelet transforms of each image. We show that by constraining the Gabor center frequencies to be low, the invariants can be optimized to capture subtle inter-species colored-texture differences. When used as input to a support vector machine classifier, the invariants provide classification of larvae to species with accuracy of 85%.