Design of an Embedded Detection System Based -Convolutional Neural Network for Swine Status and Health Assessment

Authors

  • Ryann A. Alimuin
  • Elmer P. Dadios
  • Shearyl U. Arenas

Abstract

The internal state of animals can be expressed through their behavior—changes in such behaviors can possibly be used to assess immediate signs of problems, such as animal well-being, more particularly in livestock. Appetite, movement, and activeness are factors that can be considered in assessing the swine status. Intrinsically, technological advancement in computer vision based on these behaviors are quite limited especially in open field monitoring. This paper presents the design of a Swine Status Assessment utilizing an embedded detecting system based on Convolutional Neural Network that classify each pig, record real-time activities and determine the swine’s status. Using this kind of computer vision captured from real-time video acquisition and object-oriented programming, the implementation of the design seeks to create a monitoring system which will assess the health of the pig by monitoring their behavior within specified parameters. The design is deemed practical commercial settings since it renders continuous observation of behavior with reference to several pigs at a time to validate livestock status assessment. This method will enhance the monitoring system of livestock in the swine industry in terms of their health, welfare and pig farming.

Downloads

Published

2020-01-22

Issue

Section

Articles