What question did this study set out to answer?

This research aims to enhance detection of marine wildlife trafficking using AI algorithms.

June 11, 2026Open Access

Marine wildlife trafficking: use of AI algorithms for the autodetection of shark fins, seahorses and sea cucumbers

Key Points

This research aims to enhance detection of marine wildlife trafficking using AI algorithms.
Developed algorithms using 298 scans from 68 samples of dried shark, seahorse, and sea cucumber.
Employed 3D convolutional neural networks and 3D Threat Image Projection to augment data.
Assessed detection performance and false alarm rates for target species.
Detection rates were high: shark fins 95%, seahorses 96%, sea cucumbers 86%.
Low false alarm rates were observed: 2% for shark fins, 9% for seahorses, and 1% for sea cucumbers.

Abstract

Marine wildlife trafficking is a lesser documented but significant component of wildlife trafficking. This paper expands on innovative 3D X-ray CT technology to create targeted marine wildlife algorithms using AI for the autodetection of marine wildlife in mail/air traveler pathways. A total of 298 scans derived from 68 samples of dried shark ( n = 18), dried seahorse ( n = 30) and dried sea cucumber ( n = 20) were used to develop initial algorithms. 3D convolutional neural networks and 3D Threat Image Projection were used to expand the training data set. Algorithm detection performance across all marine life subclasses was high ( P D shark fin: 95%, seahorse: 96% and sea cucumber: 86%), with low false alarm rates ( P FA shark fin: 2%, seahorse: 9% and sea cucumber: 1%). This paper demonstrates the efficacy and potential impact of AI detection tools to complement existing methods (e.g., human inspection) to combat the trafficking of target marine species globally.

Bookmark

View Full Paper