Digital imaging techniques in otolith data capture, analysis and interpretation

Research output: Contribution to journalArticle

7 Citations (Scopus)
9 Downloads (Pure)

Abstract

Otoliths or ear-stones are hard, calcium carbonate structures located within the inner ear of bony fishes. Counts of rings and measurements of seasonal growth increments from otoliths are important metrics for assessment and management of fish stocks, and the preparation and microscopic analysis of otoliths forms an essential part of the routine work undertaken by fisheries scientists worldwide. Otolith analysis is a skilled task requiring accuracy and precision, but it is laborious, time-consuming to perform, and represents a significant cost to fisheries management. In the last 2 decades, several attempts to apply ‘computer vision’ (systems that perform high-level tasks and exhibit intelligent behaviour) in otolith analysis have been reported. Although considerable progress has been made and several prototype systems developed, laboratories have been reluctant to adopt image-based computer-assisted age and growth estimation (CAAGE) systems. This paper surveys applications of CAAGE, focusing on their utility for automated ageing using images of otolith macrostructure. A cost-benefit analysis of CAAGE of cod, plaice and anchovy shows that computer vision performs relatively poorly compared with morphometric techniques. However, there is evidence that information from visual features can boost the performance of morphometric CAAGE, and further work is needed to develop effective frameworks for this integrated approach. The cost benefit of these systems might be attractive to smaller laboratories that are already using age-length keys derived from otolith morphometrics for management of smaller artisanal fisheries.
Original languageEnglish
Pages (from-to)213-231
JournalMarine Ecology Progress Series
Volume598
DOIs
Publication statusPublished - 28 Jun 2018

Keywords

  • otolith
  • computer-assisted age and growth estimation (CAAGE)
  • image analysis
  • computational model

Cite this