**1. Introduction**

In Canada, ~425,000 mature cows are harvested annually, producing over ~100,000 Tm of meat [1]. Recently, the reduced availability of cattle and the increase in beef demand have increased beef prices, particularly in cull cows [1]. In the Canadian Grading System, cull cows are segregated as Canada D-grades based on a broad classification of carcass types [2]. In contrast to the top youthful grades (Canada Prime, AAA, AA, and A), where estimations for retail cut yield are routinely provided, Canada D-grades are lacking prediction of carcass yields before carcass breakdown. Because mature beef carcasses are often boned out for further processing, yield assessments of carcasses would be an important attribute to enhance fair compensation to the cattle producers. Furthermore, accurate estimations of carcass composition have been suggested to assure an efficient utilization of specific muscles from cull cow carcasses. In this sense, Roberts et al. [3] reported that, despite darker lean, many muscles from D-grade carcasses had higher intramuscular fat content than in the youthful A/AA carcasses. Given this retail performance of muscles from cull cow carcasses, opportunities may exist to better utilize specific muscles from these carcasses.

For decades, in North America, the carcass classification has been carried out by trained personnel (graders), thus implying a certain degree of subjectivity on the quantified parameters [4]. The latest improvements in technologies to estimate body/carcass composition have shown applicability on different species, genetics, production systems, etc. [5,6].

**Citation:** Segura, J.; Aalhus, J.L.; Prieto, N.; Larsen, I.L.; Juárez, M.; López-Campos, Ó. Carcass and Primal Composition Predictions Using Camera Vision Systems (CVS) and Dual-Energy X-ray Absorptiometry (DXA) Technologies on Mature Cows. *Foods* **2021**, *10*, 1118. https://doi.org/10.3390/foods10051118

Academic Editors: Huerta-Leidenz Nelson and Markus F. Miller

Received: 28 March 2021 Accepted: 13 May 2021 Published: 18 May 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Computer vision systems (CVS) were implemented in the early 1980s as a computerized, non-destructive, non-invasive, objective, cost-effective, and automatable technology, based on image analysis that provides measurements of the beef carcass or rib-eye proportions [6]. The CVS have been recognized as useful tools to improve the grading accuracy, precision, and consistency, thus benefiting all segments of the beef production and consumption supply chain [7]. Typically, at least one of the two CVS approaches is used. Whole-side carcass image analysis, also known as hot carcass camera (HCC) system, which is designed to be integrated into the slaughter chain to work autonomously, and/or the rib-surface image analysis system, also known as cold carcass camera (CCC), which mimics the traditional visual assessment of the knife-ribbed surface of the rib-eye at the 12th thoracic vertebrae. The HCC uses a color camera and a lighting system, including structured (striped) light. The half carcass holds steady in front of a colored background and one or two images (if ambient light must be compensated) are taken to obtain 2D information, and a third image is taken with the structured light to capture 3D information of the carcass from the degree of curvature of the striped light [8,9]. Using proprietary software, the CCC provides an objective measure of rib-eye length, width, and area, and fat thickness, which are then used to predict carcass yield, as well as marbling, lean, and fat contents, and color assessments [6]. Currently, the CCC system is widely utilized by the beef industry in North America [8–11], particularly in youthful carcasses. However, unlike youthful beef, mature cull cows are generally marketed without knife-ribbing the carcass at the grade site. Hence, prediction of lean yield using rib-eye assessment or CCC is not achievable and development of alternative methods is particularly pertinent for the industry.

On the other hand, Dual-energy X-ray absorptiometry (DXA) technology is a promising indirect method to estimate carcass composition due to its relatively low cost, high reliability of data collection, and ease of use [5]. In the literature, the feasibility, accuracy, and precision of DXA technology has been reported on salmon [12], broiler chickens [13], sheep [14], swine [15], and cattle [16]. In addition, Soladoye et al. [15], Kipper et al. [17], and <sup>L</sup>ópez-Campos et al. [18] assessed the accuracy of DXA technology on mass measurement of primal cuts from pigs and steers. Most of the published studies reported on the use of DXA in youthful populations, with information being scarce or almost lacking for mature animals, particularly in the case of cull cows. Contrary to the CVS, DXA technology is at the early stages of industry implementation.

Thus, the objective of the present study was to evaluate the potential of computer vision systems, namely the whole-side carcass camera compared to the rib-eye camera, as well as the emerging DXA technology to predict whole-carcass and primal composition (fat, lean, and bone) of mature cows. Furthermore, the combination of both computer vision systems was also explored in order to evaluate this approach as an alternative for the beef industry to further improve the prediction accuracy on primals and carcass composition of mature beef.

### **2. Materials and Methods**
