We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
CONTEXT-BASED SEGMENTATION OF THE LONGISSIMUS MUSCLE IN BEEF WITH A DEEP NEURAL NETWORK.
- Authors
Talacha, Karol; Swiderski, Bartosz; Kurek, Jarosław; Kruk, Michał; Półtorak, Andrzej; Chmielewski, Leszek J.; Wieczorek, Grzegorz; Antoniuk, Izabella; Pach, Jakub; Orłowski, Arkadiusz
- Abstract
The problem of segmenting the cross-section through the longissimus muscle in beef carcasses with computer vision methods was investigated. The available data were 111 images of crosssections coming from 28 cows (typically four images per cow). Training data were the pixels of the muscles, marked manually. The AlexNet deep convolutional neural network was used as the classifier, and single pixels were the classified objects. Each pixel was presented to the network together with its small circular neighbourhood, and with its context represented by the further neighbourhood, darkened by halving the image intensity. The average classification accuracy was 96%. The accuracy without darkening the context was found to be smaller, with a small but statistically significant difference. The segmentation of the longissimus muscle is the introductory stage for the next steps of assessing the quality of beef for the alimentary purposes.
- Subjects
MUSCLE physiology; IMAGE segmentation; ARTIFICIAL neural networks; DEEP learning; ACCURACY
- Publication
Machine Graphics & Vision, 2019, Vol 28, Issue 1-4, p47
- ISSN
1230-0535
- Publication type
Article
- DOI
10.22630/mgv.2019.28.1.5