Page 17 - TAF Promo v3

Science & Innovation 17

When an animal at slaughter does not fall within shapes,” he says. “Humans do not perceive
the industry grid and Meat Standards Australia dimensions accurately, but they are inherently
(MSA) grading for eating quality, the loss for non- capable of discriminating shape differences.
compliance can be as high as $300 per beast.
“The next step was to turn a shape, such as
“At the moment we rely on poor information from a muscling, into a mathematical description and
range of different measures, and accuracy can be assign it a value.”
as low as 20 to 30 per cent. This technology could
double that accuracy rate and that would mean a Using the mathematical description along with as
transformative shift in livestock management. a fat measurement obtained by ultrasound and
the muscle score ascribed by an expert assessor,
“A novel project like this fits with our commitment Dr Alempijevic and his team taught their machine
to Australian red meat producers to support good to estimate an animal’s condition based on the 3D
science that will improve production efficiency and shape the machine senses.
create better food products for our markets.”
Dr Alempijevic used cattle assessors’ expertise in “Essentially, we are enabling computers to think
visually grading fat and muscle as the starting point and reason about what they see,” he says.
for developing his live animal imaging device.
The start of Dr Alempijevic’s research coincided
“An animal deposits muscle and fat in different with the end of a research project run by the
areas of the body, which results in different NSW Department of Primary Industries (DPI) that

>>> Back to table of contents