Automated Artemia length measurement using U-shaped fully convolutional networks and second-order anisotropic Gaussian kernels

The brine shrimp Artemia, a small crustacean zooplankton organism, is universally used as live prey for larval fish and shrimps in aquaculture. In Artemia studies, it would be highly desired to have access to automated techniques to obtain the length information from Anemia images. However, this problem has so far not been addressed in literature. Moreover, conventional image-based length measurement approaches cannot be readily transferred to measure the Artemia length, due to the distortion of non-rigid bodies, the variation over growth stages and the interference from the antennae and other appendages. To address this problem, we compile a dataset containing 250 images as well as the corresponding label maps of length measuring lines. We propose an automated Anemia length measurement method using U-shaped fully convolutional networks (UNet) and second-order anisotropic Gaussian kernels. For a given Artemia image, the designed UNet model is used to extract a length measuring line structure, and, subsequently, the second-order Gaussian kernels are employed to transform the length measuring line structure into a thin measuring line. For comparison, we also follow conventional fish length measurement approaches and develop a non-learning-based method using mathematical morphology and polynomial curve fitting. We evaluate the proposed method and the competing methods on 100 test images taken from the dataset compiled. Experimental results show that the proposed method can accurately measure the length of Artemia objects in images, obtaining a mean absolute percentage error of 1.16%.