Abstract:
This study reports on 3-dimensional (3D) imaging technologies that reconstruct the shape of both live cattle and beef carcasses. The 3D imaging technology for live cattle assesses hip height (cm), fat depth (mm) and muscle score and for beef carcasses assesses lean meat yield (%). A supervised machine learning approach using non-linear regression algorithms have been developed to assess the traits on both live cattle and carcasses. To develop the algorithms observed values are compared with the 3D assessments. The observed live cattle assessments are measured hip height, ultrasound scanning of fat depth and objective assessment of muscle score by accredited assessors; and the observed carcass assessment is computed tomography (CT) scanning of boned out primals to estimate lean and fat tissue. Cognitive perception of humans supports the view that assessors leverage shape rather than distance (e.g., when ascertaining a visual muscle score of beef cattle). Therefore, it is advantageous to represent the body shape beyond 2D and leverage 3D shape representations as curvature, which is one of the core novelties of the research. An example of predicting carcass traits from live cattle assessments before slaughter will be presented.

Biography:
Dr Malcolm McPhee began working for NSW Department of Primary Industries in November 1991 at the Elizabeth Macarthur Agricultural Institute within the Nutrition and Physiology Laboratory, Camden, Australia. He was transferred to Armidale, NSW, Australia in 1997 to work within the Beef Industry Centre of Excellence and on projects within the Beef Cooperative Research Centre. Dr McPhee’s research is focussed on assisting beef producers ‘meet market specifications’ and improve productivity and profitability. Drawing on over nearly three decades of scientific research Dr McPhee has been instrumental in the development of 4 major decision support tools for sheep and cattle producers. More recently, Dr McPhee has worked with a team of scientists in the development of the BeefSpecs calculator (http://beefspecs.agriculture.nsw.gov.au/) and BeefSpecs drafting tool (http://beefspecs.agriculture.nsw.gov.au/drafting/). The work conducted on developing BeefSpecs has led to the development of 3D imaging technologies in collaboration with the University of Technology, Sydney, Australia with funding from Meat and Livestock Australia to objectively assess hip height, P8 fat, and muscle score on live cattle. Dr McPhee has a Bachelor of Applied Science in mathematics majoring in statistics from the University of Technology, Sydney, Australia, Masters Degree in statistics from the University of Wollongong, Wollongong, Australia and a PhD in Nutritional Biology, majoring in Animal Systems modelling, from the University of California, Davis, California, USA. Dr McPhee is committed to developing technologies and decision support tools to improve the productivity and profitability of agricultural industries.

X