What is Yasiru Rangana's PhD research on?

Small object detection in fast-moving video, using convolutional neural networks including YOLO, Faster R-CNN and MobileNet, at Western Sydney University's School of Computer, Data and Mathematical Sciences.

What was Yasiru Rangana's first computer vision paper?

'Implementation of Content-Based Image Indexing and Retrieving Algorithm', with A T Dharmaratne and H A S G Dharmarathne, at ICMV 2010 in Hong Kong, from his undergraduate thesis at the University of Colombo.

Research

Yasiru Rangana is a doctoral researcher at Western Sydney University, in the School of Computer, Data and Mathematical Sciences. His research is on small object detection in fast-moving video, using convolutional neural networks including YOLO, Faster R-CNN and MobileNet. Current work is moving the detectors toward edge deployment on NVIDIA Jetson hardware.

Publications

Mitigating Roadkill Using Computer Vision and Machine Learning Techniques: A Case Study on Turtles in Australia

Yasiru Kirindearachchi and Liwan Liyanage. 2025 International Conference on Activity and Behavior Computing (ABC 2025), Khalifa University, Abu Dhabi, UAE, 21 to 25 April 2025. IEEE. DOI: 10.1109/ABC64332.2025.11118553

Roadkill surveys tend to count large animals and rely on citizen reports of remains, so small species such as turtles are under-counted. The paper trains YOLO object detectors to spot turtles on roadsides in real time, reaching a mean average precision above 90% with data augmentation and careful validation. The aim is automated roadside surveys and, eventually, systems that help prevent collisions rather than record them.

Implementation of Content-Based Image Indexing and Retrieving Algorithm

Y R Kirindearachchi, A T Dharmaratne and H A S G Dharmarathne. 3rd International Conference on Machine Vision (ICMV 2010), Hong Kong. DOI: 10.13140/RG.2.1.1321.0640

From his undergraduate thesis at the University of Colombo. The system indexes images by colour, texture and shape instead of text tags, and matches them by statistical distance. It found duplicate and resized copies with over 99% accuracy, and similar images with about 98%.

Awards

People’s Choice Award, Research Theme Image Competition, Western Sydney University Research Theme Program, November 2023, for the image “Preserving Biodiversity: AI Mitigates Roadkill for Endangered Species” (Environment and Sustainability theme). Listed in the Research Theme Program annual report 2022 to 2023.

Peer review

Panel of Reviewers, International Conference in Data Science (ICDS 2023 and 2025), Centre for Data Science, University of Colombo.

Research profiles

Questions

What is Yasiru Rangana's PhD research on?: Small object detection in fast-moving video, using convolutional neural networks including YOLO, Faster R-CNN and MobileNet, at Western Sydney University's School of Computer, Data and Mathematical Sciences.
Who wrote 'Mitigating Roadkill Using Computer Vision and Machine Learning Techniques: A Case Study on Turtles in Australia'?: Yasiru Kirindearachchi (Yasiru Rangana) and Liwan Liyanage, Western Sydney University. It was published by IEEE at ABC 2025, DOI 10.1109/ABC64332.2025.11118553.
What was Yasiru Rangana's first computer vision paper?: 'Implementation of Content-Based Image Indexing and Retrieving Algorithm', with A T Dharmaratne and H A S G Dharmarathne, at ICMV 2010 in Hong Kong, from his undergraduate thesis at the University of Colombo.

Updated 18 October 2026