Comparison of Machine Learning Algorithms for Estimation of Pig Body Weights from On-Animal and Digital Image Measurements - Thesis

Aug 6, 2026

Upload by Hazel

Live body weight (LBW) serves as a key metric for animal health and livestock market needs. This thesis evaluates the feasibility of a semiautomatic system to estimate pig LBW using four machine learning models—Linear Regression, K-Nearest Neighbor, Decision Tree, and Random Forest—based on biometric and morphometric measurements extracted from digital images. Data were collected from 26 pigs at Arkell Swine Research Facility using a consumer camera and a reference object. Age and morphometrics significantly affected predictions, while gender did not. Image-based measurements correlated positively with manual measures and achieved prediction errors within ±3 kg, with image estimates up to 16% lower than manual values, enabling cost- and time-efficient support for animal research and production.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages165

Reading Time6 h 56 min

AccessFree to download

Abstract

Acknowledgements

Table of Contents

List of Tables

List of Figures

List of Equations

List of Symbols, Abbreviations or Nomenclature

Chapter 1: General Introduction

Background of the Study

Literature Review

Study Objectives

Data Acquisition

Data Curation

Image Processing Using ImageJ

Datasets

Comparison of machine learning algorithms for estimation of pig body weights from on-animal and digital image measurements by Zhuoyi Wang A Thesis presented to The University of Guelph In partial fulfilment of requirements for the degree of Master of Science in Animal Biosciences Guelph, Ontario, Canada © Zhuoyi Wang, May, 2023 ABSTRACT COMPARISON OF MACHINE LEARNING ALGORITHMS FOR ESTIMATION OF PIG BODY WEIGHTS FROM ON-ANIMAL AND DIGITAL IMAGE MEASUREMENTS Zhuoyi Wang University of Guelph, 2023 Advisor(s): Dr. Dan Tulpan Live body weight (LBW) is an important parameter for both animal health and market requirements for livestock. My thesis focused on exploring the feasibility of a semiautomatic system that estimated the LBW of pigs by applying 4 machine learning (ML) methods (Linear Regression, K-Nearest Neighbor, Decision Tree, and Random Forest) using 3 biometric and 6 morphometric measurements extracted from digital images acquired from 26 pigs in Arkell Swine Research Facility with a consumer-level camera and a reference object. Age, age group, and 6 morphometric measurements significantly influenced LBW estimation, while gender did not. The correlation between manual and image-based measurements was positive and moderately high. Prediction errors for ML models trained on the two acquisition methods were within a ± 3 kg range, with image-based measurements performing up to 16% lower than manual measurements. This cost-and time-efficient system shows potential for intelligent solutions in animal research and production. iii ACKNOWLEDGEMENTS First and foremost, I would like to express my sincere gratitude to my supervisor, Dr. Dan Tulpan, for seeing potential in me since I was a 4th-year undergraduate student and for taking me as graduate student. Thank you so much for trusting me, giving me the opportunity, and guiding me to explore my potential. You always share your experience to teach me and use your sense of humor to encourage me. Your patience, support, and advice have lighted my road both in academic study and in daily life. Especially being an international student living in this foreign country, I felt the warmness from you and all the members in our lab. Everything I have learnt from you was invaluable for my future life. I am deeply grateful to my advisory committee, Dr. Renee Bergeron, and Dr. Christine Baes, for supporting me with the data collection process and giving me advice on the experimental design and the thesis from different perspectives. I have learnt a lot from each of you through all the stages of this project. Thank you to Jasmine Muszik, Troy McElwain, Douglas Wey, Jonathan Duncan and all the other staff working at the Arkell Swine Research Facility for your help with image capturing, body measuring, and LBW weighing. I appreciate all of your active help eventhough we were under such a hard time with many restrictions due to the COVID-19 pandemic. I would like to say thank you to all my lab mates, Esther Chan, Kaitlyn Rodriguez, Jihao You, and Saeed Shadpour. I am so glad to have you all here working hard together as comrade in arms. We learned together, grew together, and supported each other for two years. This will be one of the most cherished parts of my life. Finally, I want to thank my family for their support providing me the chance and confidence to study abroad. I am also grateful to have my boyfriend with me for the past nine years helping me release stress and using his silent love to be my battery booster. iv TABLE OF CONTENTS Abstract .................................................................................................................................. ii Acknowledgements ............................................................................................................... iii Table of Contents ..............................................................................................................

FAQs

What problem does the thesis address?

The thesis addresses estimating pigs’ live body weight (LBW) accurately for both animal health and market requirements using measurements from digital images.

Which machine learning methods are compared?

Four methods are compared: Linear Regression, K-Nearest Neighbor, Decision Tree, and Random Forest.

How does image-based measurement performance compare with manual measurement?

Prediction errors for models trained on manual vs. image-based data stay within a ±3 kg range, and image-based measurements can be up to 16% lower than manual measurements.

Aug 6, 2026

Upload by Hazel