Date Published: December 17, 2018
Publisher: Public Library of Science
Author(s): Endre Moen, Nils Olav Handegard, Vaneeda Allken, Ole Thomas Albert, Alf Harbitz, Ketil Malde, Heather M. Patterson.
The age structure of a fish population has important implications for recruitment processes and population fluctuations, and is a key input to fisheries-assessment models. The current method of determining age structure relies on manually reading age from otoliths, and the process is labor intensive and dependent on specialist expertise. Recent advances in machine learning have provided methods that have been remarkably successful in a variety of settings, with potential to automate analysis that previously required manual curation. Machine learning models have previously been successfully applied to object recognition and similar image analysis tasks. Here we investigate whether deep learning models can also be used for estimating the age of otoliths from images. We adapt a pre-trained convolutional neural network designed for object recognition, to estimate the age of fish from otolith images. The model is trained and validated on a large collection of images of Greenland halibut otoliths. We show that the model works well, and that its precision is comparable to documented precision obtained by human experts. Automating this analysis may help to improve consistency, lower cost, and increase the extent of age estimation. Given that adequate data are available, this method could also be used to estimate age of other species using images of otoliths or fish scales.
Age of fish is a key parameter in age-structured fisheries-assessment models. Age is usually considered as a discrete parameter (age group) that identifies the individual year class i.e. those originating from the spawning activity in a given year . An individual is categorized as age group 0 from the first early larval stage, and all age groups increase their age at 1 January. Assessment models typically express the dynamics of the individual year class from the age when they recruit, through sexual maturation, reproduction, and throughout their life cycle . The models are fitted to data originating from commercial catches and fisheries-independent surveys. A sampling program for a specific fish stock typically involves sampling throughout the year using several different types of fishing gears.
Predictions were made on the test set for the different configurations and the MSEs of single otolith predictions of age were recorded. The MSE and CV of pair-wise predictions were also recorded. Calculated CV values were then used to select the optimum CNN model.
The objective of this study was to investigate to what extent a deep CNN could be adapted to predict age from otolith images. Using a data set of Greenland halibut otoliths, we trained and validated an Inception-3 network and showed that it performed at a level close to human accuracy. Deep neural networks have been shown to outperform more conventional methods across a range of problems , and given their generality, we hypothesized that they would perform well on this rather difficult task. Several different network architectures were used, and most configurations were able to perform well, which further supports our hypothesis.