Genetic parameters for image-based estimations of swine feet and leg conformation traits.

IF 2.7 2区 农林科学 Q1 AGRICULTURE, DAIRY & ANIMAL SCIENCE
Zack C Peppmeier, Yijian Huang, Jan-Marie B Bartholomew, Jicai Jiang, Mark T Knauer, Suzanne M Leonard
{"title":"Genetic parameters for image-based estimations of swine feet and leg conformation traits.","authors":"Zack C Peppmeier, Yijian Huang, Jan-Marie B Bartholomew, Jicai Jiang, Mark T Knauer, Suzanne M Leonard","doi":"10.1093/jas/skaf103","DOIUrl":null,"url":null,"abstract":"<p><p>The objectives of this study were to develop and evaluate a novel algorithm for image extraction of structural conformation traits and estimate variance components among skeletal conformation, growth, and herd retention traits. An Intel RealSense D435i camera was used to obtain left side-view RGB images on individual purebred Duroc pigs (n = 846) at 156 d of age. Frames were selected by a trained swine evaluator when either the left front leg (n = 1056), left back leg (n = 888), or both left legs (n = 728) were present in the field of view and the respective foot pads from toe to heel were in contact with the ground. Selected images were processed through Apple Inc's image segmentation algorithm to extract the pig from the background. Segmented pig images were then processed through a novel algorithm developed in this study. The algorithm identified the leg and estimated 21 skeletal conformation traits from each leg. Steps for user intervention were added to assist the algorithm in identifying which leg(s) were present and the general location of each leg to increase accuracy of leg identification and trait acquisition. The algorithm correctly identified at least one front and one back leg from an image for 99.9% and 98.0% of the pigs, respectively. Heritability estimates ranged from 0.01 to 0.33 for all conformation traits with the quadratic term for the curvature of the anterior side of the front and the height of the back leg having the highest heritability for each location (h2 = 0.33 and 0.30, respectively). Genetic correlations among image feet and leg conformation traits and production traits (finishing average daily gain, weight per day of age, and finishing feed efficiency) ranged from -0.37 to 0.19. Boars that remained in the breeding herd for longer than 200 days tended (p = 0.08) to have greater curvature of the front leg and lower (p = 0.07) angularity between the midpoint of foot and the anterior point of the pastern and had significantly (p = 0.03) shorter distance between the pastern and the top of the shoulder than those that were removed prior to 200 days. Gilts that remained in the breeding herd for longer than 200 days tended (p = 0.08) to have less curvature of the back leg. The current study presents an algorithm that extracts novel, objective structural conformation traits and reports corresponding genetic and phenotypic parameters.</p>","PeriodicalId":14895,"journal":{"name":"Journal of animal science","volume":" ","pages":""},"PeriodicalIF":2.7000,"publicationDate":"2025-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of animal science","FirstCategoryId":"97","ListUrlMain":"https://doi.org/10.1093/jas/skaf103","RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"AGRICULTURE, DAIRY & ANIMAL SCIENCE","Score":null,"Total":0}
引用次数: 0

Abstract

The objectives of this study were to develop and evaluate a novel algorithm for image extraction of structural conformation traits and estimate variance components among skeletal conformation, growth, and herd retention traits. An Intel RealSense D435i camera was used to obtain left side-view RGB images on individual purebred Duroc pigs (n = 846) at 156 d of age. Frames were selected by a trained swine evaluator when either the left front leg (n = 1056), left back leg (n = 888), or both left legs (n = 728) were present in the field of view and the respective foot pads from toe to heel were in contact with the ground. Selected images were processed through Apple Inc's image segmentation algorithm to extract the pig from the background. Segmented pig images were then processed through a novel algorithm developed in this study. The algorithm identified the leg and estimated 21 skeletal conformation traits from each leg. Steps for user intervention were added to assist the algorithm in identifying which leg(s) were present and the general location of each leg to increase accuracy of leg identification and trait acquisition. The algorithm correctly identified at least one front and one back leg from an image for 99.9% and 98.0% of the pigs, respectively. Heritability estimates ranged from 0.01 to 0.33 for all conformation traits with the quadratic term for the curvature of the anterior side of the front and the height of the back leg having the highest heritability for each location (h2 = 0.33 and 0.30, respectively). Genetic correlations among image feet and leg conformation traits and production traits (finishing average daily gain, weight per day of age, and finishing feed efficiency) ranged from -0.37 to 0.19. Boars that remained in the breeding herd for longer than 200 days tended (p = 0.08) to have greater curvature of the front leg and lower (p = 0.07) angularity between the midpoint of foot and the anterior point of the pastern and had significantly (p = 0.03) shorter distance between the pastern and the top of the shoulder than those that were removed prior to 200 days. Gilts that remained in the breeding herd for longer than 200 days tended (p = 0.08) to have less curvature of the back leg. The current study presents an algorithm that extracts novel, objective structural conformation traits and reports corresponding genetic and phenotypic parameters.

基于图像估计猪脚和腿形态特征的遗传参数。
本研究的目的是开发和评估一种新的结构构象特征图像提取算法,并估计骨骼构象、生长和群体保留特征之间的方差成分。使用英特尔RealSense D435i相机获取156日龄杜洛克猪(n = 846)的左侧视RGB图像。当左前腿(n = 1056)、左后腿(n = 888)或两条左腿(n = 728)出现在视野中,并且各自的脚垫从脚趾到脚跟与地面接触时,由训练有素的猪评估员选择框架。选取的图像通过苹果公司的图像分割算法进行处理,从背景中提取出猪。然后,通过本研究开发的一种新算法对分割的猪图像进行处理。该算法确定了腿,并估计了每条腿的21个骨骼构象特征。增加了用户干预步骤,以帮助算法识别哪些腿存在以及每条腿的大致位置,以提高腿识别和特征获取的准确性。该算法分别为99.9%和98.0%的猪正确地从图像中识别出至少一条前腿和一条后腿。所有构象性状的遗传率估计范围在0.01 ~ 0.33之间,每个位置的前侧面曲率和后腿高度的二次项遗传率最高(h2分别= 0.33和0.30)。象脚和腿形态性状与生产性状(肥育平均日增重、日龄平均体重和肥育饲料效率)的遗传相关性在-0.37 ~ 0.19之间。在种猪群中停留超过200天的公猪,其前腿弯曲度(p = 0.08)更大,足中点与骨节前点之间的角度更低(p = 0.07),且骨节与肩顶之间的距离显著(p = 0.03)更短(p = 0.03)。在种畜群中停留超过200天的后备母猪,其后腿弯曲度更小(p = 0.08)。目前的研究提出了一种算法,提取新的,客观的结构构象特征,并报告相应的遗传和表型参数。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
Journal of animal science
Journal of animal science 农林科学-奶制品与动物科学
CiteScore
4.80
自引率
12.10%
发文量
1589
审稿时长
3 months
期刊介绍: The Journal of Animal Science (JAS) is the premier journal for animal science and serves as the leading source of new knowledge and perspective in this area. JAS publishes more than 500 fully reviewed research articles, invited reviews, technical notes, and letters to the editor each year. Articles published in JAS encompass a broad range of research topics in animal production and fundamental aspects of genetics, nutrition, physiology, and preparation and utilization of animal products. Articles typically report research with beef cattle, companion animals, goats, horses, pigs, and sheep; however, studies involving other farm animals, aquatic and wildlife species, and laboratory animal species that address fundamental questions related to livestock and companion animal biology will be considered for publication.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信