UNIMAS, SFC team up: Advancing sarawak turtle conservation

UNIMAS, SFC team up: Advancing sarawak turtle conservation is profiled by BTW Media because public-source evidence links it to internet infrastructure, governance, operational dependencies, or market visibility.

• UNIMAS and SFC collaborate on turtle re-identification.
• Development of untagged/unmarked recognition prototype.

In a pioneering effort, a collaborative research endeavour between the University Malaysia Sarawak (UNIMAS) and the Sarawak Forestry Corporation (SFC) has yielded a breakthrough in turtle conservation. Focused on female green turtles, the project aimed to revolutionize re-identification methods without relying on traditional labeling methods. Departing from conventional plastic or metal tags, the team opted for a novel approach, leveraging downward-facing cameras to capture high-resolution images of nesting individuals.

Preserving sarawak’s ecosystem

Sarawak’s rich biodiversity is under significant threat due to human activities and environmental degradation. Pollution, overfishing, and habitat destruction particularly endanger coastal areas. Pressure on coral reefs, crucial for marine biodiversity and fisheries, comes from sedimentation, chemical runoff, and coral bleaching caused by climate change. Sarawak hosts various marine species, including iconic green turtles, vital for the health of coral reefs and seagrass beds.

The loss of biodiversity in Sarawak’s ecosystems jeopardizes their stability and the services they provide, including clean air and water, food security, climate regulation, and cultural significance to indigenous communities.

Also read: UTP’s 1st neurodiversity seminar advances understanding

AI advancements in conservation

Under SFC supervision, some sites designated as Total Protection Areas (TPAs) to minimize human interaction, aiding population growth. Team explores camera-based imaging instead of traditional tagging.

UNIMAS’s Khalif Amir Zakry and Syahiran Soria develop AI methods to address turtle identification challenges. Syahiran trains a deep learning model to automatically crop shell regions of interest (ROIs).

Khalif combines trained models with attention mechanisms to guide networks in identifying stable features over time. Models trained on image datasets collected over two years from Talang-Satang Island.

Both students’ research papers accepted by SCOPUS-indexed journals, pending publication. Despite prototype improvements needed, the team contributes to protecting these magnificent creatures and their fragile ecosystems. Future work involves extending data collection periods and rigorous validation on larger datasets.

Also read: Singapore’s Financial Regulator Sets Sights on Web3 and Startup Ecosystem with S$150 Million Investment