Image Segmentation of Meliponine Bee using Faster R-CNN


Mohd Isa, Wan Noorshahida and Nizam, Afif and Ali, Aziah (2019) Image Segmentation of Meliponine Bee using Faster R-CNN. In: 2019 Third World Conference on Smart Trends in Systems Security and Sustainablity (WorldS4), 30-31 July 2019, London, UK.

[img] Text
104.pdf - Published Version
Restricted to Repository staff only

Download (2MB)


There has been an increasing interest by local Malaysian towards beekeeping of Meliponine, which is a tribe of stingless bee. When it comes to advice on beekeeping, these locals (mostly novice) depends on inputs from only a handful of key experts (mostly aging). This paper presents part of a project that aims to develop a visual-based expert system, which may help in beekeeping tasks. One of the first processes in such system is image segmentation. Being small in size, the segmentation of a Meliponine in its natural surrounding is difficult to be accurately detected, more so identified. Of recent, the Convolutional Neural Network (CNN) has been making tremendous progress in object detection due to its high accuracy and fast execution. In this paper, the Faster R-CNN, an object detection method, which uses the CNN core module is implemented for segmentation of Meliponine image from its background. We present results of this implementation on our data set of 400 image frames from videos collected in a local Meliponine farm in Malaysia. The accuracy result at around 74% looks promising for further exploration.

Item Type: Conference or Workshop Item (Paper)
Uncontrolled Keywords: Image segmentation, deep learning, object detection, meliponine, Malaysia
Subjects: Q Science > QP Physiology > QP351 Neurophysiology and Neuropsychology
Divisions: Faculty of Computing and Informatics (FCI)
Depositing User: Ms Rosnani Abd Wahab
Date Deposited: 27 Oct 2021 04:56
Last Modified: 27 Oct 2021 04:56


Downloads per month over past year

View ItemEdit (login required)