Detection of endangered species proboscis monkey using computer vision technique on low compute device

Juanrico Alvaro, Gede Putra Kusuma

Abstract


The uncommon and endemic proboscis monkey, also known as Nasalis larvatus, is a unique member of Indonesia's diverse animal population that is in the list of endangered species. With the development of technology, the ability to detect this species can be proven useful to further look out this species on the dense forest, their inhabitant. This research paper presents the dataset and compares the performance and computational load to run on a low computation device from three state of the art object detection models: You Only Look Once version 7 (YOLOv7), Single Shot Multibox Detection (SSD) and Faster Region Convolutional Neural Network (Faster R-CNN). All the models are evaluated using low computing device with their respective backbone MobileNetV2 for SSD and tiny-version for YOLOv7. According to the report, Faster R-CNN give the best accuracy of 89.16 and 53.46 in AP0.5 and AP[0.5:0.95] on test dataset with the slowest time among other models with FPS of 1.26 s and huge memory usage. In the other hand, SSD give the fastest detection speed with FPS of 0.31s and the lightest model to run in low computing devices with a decent average precision of 81.35 and 42.99 in AP0.5 and AP[0.5:0.95] respectively. Hence, the best proposed model for real-time detection is SSD with MobileNetV2 as the backbone.

Full Text: PDF

Published: 2023-12-29

How to Cite this Article:

Juanrico Alvaro, Gede Putra Kusuma, Detection of endangered species proboscis monkey using computer vision technique on low compute device, Commun. Math. Biol. Neurosci., 2023 (2023), Article ID 136

Copyright © 2023 Juanrico Alvaro, Gede Putra Kusuma. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Commun. Math. Biol. Neurosci.

ISSN 2052-2541

Editorial Office: [email protected]

 

Copyright ©2024 CMBN