Image
SDG14
LIFE BELOW WATER
14.5  Maintaining a local ecosystem
14.5.2  Monitoring the health of aquatic ecosystems

Monitor the health of aquatic ecosystems
No.
Description
Evidence
1.

A joint research team from Universiti Malaysia Sarawak (UNIMAS) and Sarawak Forestry Corporation is pioneering a new, stress-free method to identify individual green sea turtles (Chelonia mydas) that nest on Sarawak’s remote beaches — all without using traditional tags. Instead of plastic or metal markers, the team uses a downward-facing camera to capture images of the turtle carapace, and advanced artificial intelligence (AI) model to recognize each turtle by its unique scute (shell) patterns, similar to fingerprints. Two MSc students from UNIMAS developed this deep-learning system: one model crops the carapace automatically, while another, equipped with an attention mechanism, identifies stable visual features over time, even as the turtle ages or its shell changes. The dataset for training the AI was collected over two years on Talang‑Satang island. The approach reduces stress on turtles, avoids invasive tagging, and supports long-term monitoring; though still in prototype stage, the team plans to expand data collection and validate the system with larger datasets

2.

On 17 October 2024, The Baleh Book was officially presented to YBhg. Prof. Dr. Hj. Ahmad Hata bin Rasit, Vice-Chancellor of Universiti Malaysia Sarawak (UNIMAS), in a ceremony attended by the Director of UNIMAS Publisher. The Institute of Biodiversity and Environmental Conservation (IBEC) also had the honour of presenting the book to YBhg. Tuan Abang Arabi Abang Aimran, General Manager of Sarawak Forestry Corporation. This publication is the outcome of a collaborative effort with Sarawak Energy Berhad, which provided a research grant to UNIMAS to explore the rich biodiversity of the “lost world” of Baleh. Sincere appreciation is extended to all reviewers and authors for their invaluable contributions to this significant work.

Download UNIMAS Now App

Image
Image