New Imaging Technologies Could Transform Wagyu Carcase Yield Prediction

Emerging 3D imaging and artificial intelligence technologies could revolutionise the way the beef industry measures carcase yield, providing producers, feedlots and processors with more accurate data than ever before.

That was the message from Cooper Carter, recipient of an Australian Wagyu Association Fellowship and Masters student at Texas Tech University, who presented his latest research at WagyuEdge’27 in April. Cooper is investigating how advanced imaging technologies can be used to more accurately predict retail meat yield in Wagyu and beef carcases.

Speaking to delegates, Cooper outlined how current yield grading systems were developed using measurements and production systems that look very different from those used today. While traditional grading methods provide an indication of expected yield, he believes modern technology offers an opportunity to significantly improve accuracy and deliver more meaningful information back to producers.

Rethinking Traditional Yield Prediction

Much of Cooper’s research focuses on the United States Department of Agriculture (USDA) yield grading system, which predicts the percentage of boneless, closely trimmed retail cuts from a carcase. The system relies primarily on eye muscle area, fat measurements and hot carcase weight.

However, Cooper highlighted research showing that while the system identifies broad trends across populations, it can be considerably less accurate when predicting yield on an individual animal basis. Research conducted at Texas Tech University demonstrated that some carcases graded as lower yielding actually outperformed those graded as higher yielding when true retail meat yield was measured.

This disconnect, he explained, comes from relying on a limited number of measurements to predict the composition of an entire carcase.

“We’re really seeing the limitations of using single-point measurements on a carcase to predict total red meat yield,” Cooper said.

Bigger Carcases, New Challenges

Another challenge facing existing grading systems is the ongoing increase in carcase weights.

Over recent decades, both the US and Australian beef industries have seen steady growth in average carcase size as producers work to improve efficiency and increase beef production. Yet heavier carcases do not always translate to better yields if additional weight is being deposited as fat rather than muscle.

Understanding the point at which animals stop efficiently producing muscle and begin depositing excess fat remains a key challenge for the industry.

“What’s the ideal weight to produce cattle to while maintaining high yield percentages?” Cooper asked delegates.

Looking Beyond the Eye Muscle

One of the key findings highlighted during the presentation was that eye muscle area alone may not be telling the complete story.

Research conducted on beef-on-dairy cattle in the United States found substantial variation in overall carcase muscling despite animals sharing similar genetics and exhibiting similar eye muscle measurements.

The findings suggest that while breeders have successfully selected for larger eye muscle area, there may be opportunities to capture additional information about muscling throughout the entire carcase, particularly in the hindquarter and forequarter.

“Rib eye area is very good for predicting some of the major cuts, but it doesn’t necessarily tell us what the rest of the carcase looks like,” Cooper explained.

The Potential of 3D Imaging

At the centre of Cooper’s research is the use of 3D imaging technology.

Using light detection and ranging (LiDAR) technology, similar to systems used in mapping and autonomous navigation, researchers can create highly detailed three-dimensional models of beef carcases. These models allow measurements of volume, shape, cross-sectional area and overall carcase conformation.

The technology is non-invasive and can be applied without physically altering the carcase, making it attractive for commercial processing environments.

Initial pilot studies involving 3D imaging and machine learning have already demonstrated promising results. While early research involved relatively small sample sizes, Cooper said expanding datasets and advances in artificial intelligence are expected to significantly improve predictive accuracy.

Researchers at Texas Tech have since expanded their database from an initial 40 head to approximately 500 cattle across multiple processing plants. Cooper said the goal is to achieve prediction accuracies well into the 90% range.

AI Unlocking New Opportunities

A major driver behind the technology’s potential is artificial intelligence.

Machine learning systems can analyse enormous numbers of data points generated from 3D scans and identify patterns that would be impossible to process manually. Researchers are also using data augmentation techniques, allowing algorithms to create additional virtual carcase models to strengthen predictive capability.

“AI is really what unlocks the power of this technology,” Cooper said.

Building the Future with CT Scanning

Alongside 3D imaging, Cooper discussed the use of computed tomography (CT) scanning, widely regarded as the gold standard for accurately determining muscle, fat and bone composition.

Researchers at Texas Tech have developed a mobile CT scanning unit that can be deployed at processing facilities. While the technology is not intended for commercial chain-speed grading, it provides highly accurate reference measurements that can be used to validate and refine predictive imaging systems.

The technology also creates exciting opportunities for future Wagyu research, particularly in evaluating intramuscular fat distribution and understanding the unique composition of Wagyu carcases.

Wagyu Collaboration Underway

As part of his Australian Wagyu Association Fellowship, Carter has already begun working directly with Australian Wagyu carcases.

During a recent visit to Australia, he completed 3D scans on approximately 130 Wagyu carcases, creating a valuable dataset that will contribute to future analysis and model development.

The long-term vision is to generate accurate red meat yield predictions that can be returned to producers and incorporated into breeding programs through future Wagyu Breeding Values (WBVs).

Such a tool could provide Wagyu breeders with a powerful new way to select cattle that not only excel in eating quality but also maximise carcase yield and production efficiency

Benefits Across the Supply Chain

Cooper emphasised that the impact of the technology extends well beyond seedstock production.

More accurate yield data could help feedlots better identify optimal feedlot exit points, assist processors in predicting carcase performance and improve the flow of information throughout the entire supply chain.

Ultimately, he believes technology-driven yield prediction has the potential to fundamentally reshape how beef is valued and produced.

“This could revolutionise how we produce beef cattle in both the United States and Australia,” Carter said.

For the Wagyu sector, the work represents an exciting step toward combining advanced technology, artificial intelligence and precision breeding to deliver greater value throughout the supply chain.

IMPORTANT NOTICE - Terms & Conditions


It is a condition of use of this database that you accept these terms and conditions. 

It is important that you appreciate when accessing the Australian Wagyu Association Limited (AWA) database (which may be hosted on an external website) that the information contained on the AWA database, including but not limited to pedigree, DNA information, Wagyu Breeding Values (WBVs) and Index values (together called Database Information), is based on data supplied by AWA members and/or third parties. 

Whilst sincere effort is made to ensure the Database Information is accurate and complete, to the extent permitted by law, AWA, its officers and employees, shall not be responsible for its use or interpretation, and you use the Database Information at your own risk. Please be aware of the following limitations of the Database Information.  

Regarding WBVs and Index values, it is important to appreciate, and you need to be aware that: 

  • WBVs are derived using genetic evaluation technology developed for the AWA, using information contained within the AWA database and parameters derived from it.
  • WBVs are estimates of genetic potential of individual animals and may change with addition of new data or with new parameters.
  • WBVs are genetic merit estimates only and may not reflect the raw animal phenotype.
  • WBVs can only be directly compared to other WBVs calculated in the same AWA genetic analysis.

Regarding pedigree and DNA testing results submitted to the AWA, it is important to appreciate, and you need to be aware that: 

  • Animals whose registration identifiers start with PED are Pedigree Recorded Animals. What this means is that the details which AWA used to register the animal were obtained from a Recognised Wagyu Registry Organisation – in other words a Wagyu breed association based outside of Australia. AWA has not independently verified the information with respect to the animal or the details recorded about the animal. You should read AWA’s by-laws which set out details regarding the definition of a Pedigree Recorded Animal. 
  • Pedigree and DNA data submitted and supplied to AWA may have errors in it which cannot be detected without further DNA testing. 
  • Technology may have advanced since a particular test was undertaken so that genetic issues or inaccuracies which were previously not detectable are now able to be detected by current testing technology (i.e. if the animal was tested again with current technology). 
  • AWA estimates that less than 1% of the pedigree entries, ownership or breeding details in the AWA Herdbook may have errors or which may be misleading. For this reason, users ought to consider if they need to obtain independent testing of the relevant animal (if possible) to ensure that the data is accurate. 

Regarding prefectural content, it is important to appreciate, and you need to be aware that: 

  • Prefectural content is based on the estimation of prefectural origin from Japanese breeding records of 201 foundation sires and 168 foundation dams. As genotype-based parent verification is not used in Japan, and full Japanese registration certificates are not available for all foundation animals, exact prefectural composition for these sires and dams cannot be validated. 
  • The calculation of prefectural content for Australian Herdbook animals relies on the accuracy of pedigree records and DNA samples provided by AWA members.   

Regarding DNA testing and genetic condition results submitted to the AWA, it is important to appreciate, and you need to be aware that: 

  • As genetic testing results are provided by independent DNA testing laboratories at the request of AWA members, AWA relies solely on the accuracy of the reporting of genetic testing as completed by the laboratories. 
  • AWA uses GeneProb software to calculate the likelihood of pedigree animals being a carrier of genetic conditions.  This likelihood is an estimate based on the available genetic testing results. 

If you consider that you do not understand or appreciate the nature and/or implications of the data provided on this website or the WBVs of a particular animal, then AWA strongly recommends that you seek expert advice.  Note there is substantial technical information and articles on the Wagyu website to assist our members. 

AWA’s liability for any loss or damage, consequential or otherwise, suffered or incurred by you or your related business, arising directly or indirectly from your use of Database Information, to the extent permitted by law, is limited to AWA providing you with a refund for the fees paid for the service (if any), or AWA re-performing the service, where possible, at the election of AWA.  

AWA administers this database in good faith to support AWA members and the industry. AWA does not represent or warrant that the animals recorded in this database, or animals breed from them, will achieve any particular genetic or phenotype performance.


THIS DISCLAIMER HAS BEEN UPDATED AS OF 03/02/2026, PLEASE ENSURE YOU HAVE READ IT IN DETAIL