Towards Automatic Cetacean Photo-Identification: A Framework for Fine-Grain, Few-Shot Learning in Marine Ecology

TL;DR

This paper introduces a fully automated computer vision framework for cetacean photo-identification that improves accuracy and efficiency by utilizing all available information and handling uncatalogued individuals in unedited field images.

Contribution

It presents a novel pipeline combining detection and matching models for automatic cetacean identification without pre-processing, capable of handling new individuals and large datasets.

Findings

Achieved high detection accuracy with mAP@IOU[0.5] of 0.91 and 0.96 on different catalogues.

Attained top-10 individual classification accuracy of 83.1% and 97.5%.

System effectively processes unedited field images and identifies uncatalogued cetaceans.

Abstract

Photo-identification (photo-id) is one of the main non-invasive capture-recapture methods utilised by marine researchers for monitoring cetacean (dolphin, whale, and porpoise) populations. This method has historically been performed manually resulting in high workload and cost due to the vast number of images collected. Recently automated aids have been developed to help speed-up photo-id, although they are often disjoint in their processing and do not utilise all available identifying information. Work presented in this paper aims to create a fully automatic photo-id aid capable of providing most likely matches based on all available information without the need for data pre-processing such as cropping. This is achieved through a pipeline of computer vision models and post-processing techniques aimed at detecting cetaceans in unedited field imagery before passing them downstream for individual level catalogue matching. The system is capable of handling previously uncatalogued individuals and flagging these for investigation thanks to catalogue similarity comparison. We evaluate the system against multiple real-life photo-id catalogues, achieving mAP@IOU[0.5] = 0.91, 0.96 for the task of dorsal fin detection on catalogues from Tanzania and the UK respectively and 83.1, 97.5% top-10 accuracy for the task of individual classification on catalogues from the UK and USA.