Investigating neural network algorithms for imaging points of interest identification

29 March 2019
Focus area: Advanced manufacturing
Program stream: Hands-off processing
Project number: 2019-1017
The objective of this project was to construct training data and create a neural network to accurately calculate beef rib scribing cut locations from X-ray images. 

Overall, the neural network greatly improved the accuracy of determining the side orientation, locating the rib 1 junction and determining the rib 2 and 8 intersections between the rib 1 junction and the Aitch bone. 
 

Supporting documents

Most recent reports

Previous in this focus area 28 February 2019 Scoping processor requirements in beef cutting control, yield data traceability, robotics, automatio Next in this focus area 10 June 2020 Validating the E+V cold carcase grading camera for Australian carcase grading