Please use this identifier to cite or link to this item:
Title: Video image analysis in the Australian meat industry - precision and accuracy of predicting lean meat yield in lamb carcasses
Contributor(s): Hopkins, DL (author); Safari, E (author); Thompson, John Mitchell  (author); Smith, CR (author)
Publication Date: 2004
DOI: 10.1016/j.meatsci.2003.10.015
Handle Link:
Abstract: A wide selection of lamb types of mixed sex (ewes and wethers) were slaughtered at a commercial abattoir and during this process images of 360 carcasses were obtained online using the VIAScan® system developed by Meat and Livestock Australia. Soft tissue depth at the GR site (thickness of tissue over the 12th rib 110 mm from the midline) was measured by an abattoir employee using the AUS-MEAT sheep probe (PGR). Another measure of this thickness was taken in the chiller using a GR knife (NGR). Each carcass was subsequently broken down to a range of trimmed boneless retail cuts and the lean meat yield determined. The current industry model for predicting meat yield uses hot carcass weight (HCW) and tissue depth at the GR site. A low level of accuracy and precision was found when HCW and PGR were used to predict lean meat yield (R²=0.19, r.s.d.=2.80%), which could be improved markedly when PGR was replaced by NGR (R²=0.41, r.s.d.=2.39%). If the GR measures were replaced by 8 VIAScan® measures then greater prediction accuracy could be achieved (R²=0.52, r.s.d.=2.17%). A similar result was achieved when the model was based on principal components (PCs) computed from the 8 VIAScan® measures (R²=0.52, r.s.d.=2.17%). The use of PCs also improved the stability of the model compared to a regression model based on HCW and NGR. The transportability of the models was tested by randomly dividing the data set and comparing coefficients and the level of accuracy and precision. Those models based on PCs were superior to those based on regression. It is demonstrated that with the appropriate modeling the VIAScan® system offers a workable method for predicting lean meat yield automatically.
Publication Type: Journal Article
Source of Publication: Meat Science, 67(2), p. 269-652
Publisher: Elsevier BV
Place of Publication: Netherlands
ISSN: 0309-1740
Field of Research (FOR): 070299 Animal Production not elsewhere classified
Socio-Economic Objective (SEO): 830310 Sheep - Meat
Peer Reviewed: Yes
HERDC Category Description: C1 Refereed Article in a Scholarly Journal
Statistics to Oct 2018: Visitors: 84
Views: 84
Downloads: 0
Appears in Collections:Journal Article

Files in This Item:
2 files
File Description SizeFormat 
Show full item record


checked on Nov 30, 2018

Page view(s)

checked on Mar 4, 2019
Google Media

Google ScholarTM



Items in Research UNE are protected by copyright, with all rights reserved, unless otherwise indicated.