Optimizing high-density aquaculture rotifer Detection using deep learning algorithm

Alixson Polumpung and Kit Guan Lim and Min Keng Tan and Sitti Raehanah Muhamad Shaleh and Renee Ka Yin Chin and Kenneth Teo Tze Kin (2022) Optimizing high-density aquaculture rotifer Detection using deep learning algorithm.

	Text ABSTRACT.pdf Download (49kB)
	Text FULL TEXT.pdf Restricted to Registered users only Download (528kB) | Request a copy

Abstract

The dynamics of marine aquaculture depend heavily on zooplankton, which is the basis of the marine food chain. Zooplankton like Rotifer brachionus plicatillis, which are rich in nutrients, small size and rapid reproductive rate are necessary for fish in the larval stage. Rotifer must therefore be supplied to larvae culture in the correct quantity, which can be determined by counting it. In addition, it is necessary to estimate the rotifer population to ensure that, aside from care, it can support the demands of all larvae batches. Currently, the traditional method of counting small-sized rotifers still involves counting it manually. One easy potential way to count rotifer is by using binary image segmentation provided that the sample is clear from debris. In this paper, we present the method and performance to detect rotifer Brachionus plicatilis in 1ml sample automatically using deep learning algorithm YOLOv3. Detected rotifer will be counted for estimating the amount of rotifer for feeding or the density population in a rotifer culture. The method of this project consists of following steps. First, dataset acquisition from digital microscope and manual labelling annotation divided by 60, 20 and 20 percent for training, validation and testing consecutively. Second, is to develop the deep learning algorithm based on YOLOv3. Third step is to training and evaluate the model using loss function. Finally, the model is tested with average precision of 85.1 percent with average of 1.4645s inference detection speed

Item Type:	Proceedings
Keyword:	deep learning, object detection, rotifer dataset
Subjects:	Q Science > QA Mathematics > QA1-939 Mathematics > QA71-90 Instruments and machines > QA75-76.95 Calculating machines S Agriculture > SH Aquaculture. Fisheries. Angling > SH1-691 Aquaculture. Fisheries. Angling > SH20.3-191 Aquaculture
Department:	FACULTY > Faculty of Engineering
Depositing User:	ABDULLAH BIN SABUDIN -
Date Deposited:	04 Nov 2024 09:40
Last Modified:	04 Nov 2024 09:40
URI:	https://eprints.ums.edu.my/id/eprint/41729

Actions (login required)

View Item