Predicting beef tenderness using color and multispectral image texture features

The objective of this study was to investigate the usefulness of raw meat surface characteristics (texture) in predicting cooked beef tenderness. Color and multispectral texture features, including 4 different wavelengths and 217 image texture features, were extracted from 2 laboratory-based multispectral camera imaging systems. Steaks were segregated into tough and tender classification groups based on Warner–Bratzler shear force. The texture features were submitted to STEPWISE multiple regression and support vector machine (SVM) analyses to establish prediction models for beef tenderness. A subsample (80%) of tender or tough classified steaks were used to train models which were then validated on the remaining (20%) test steaks. For color images, the SVM model correctly identified tender steaks with 100% accurately while the STEPWISE equation identified 94.9% of the tender steaks correctly. For multispectral images, the SVM model predicted 91% and STEPWISE predicted 87% average accuracy of beef tender.     

Evaluation of a probe for predicting beef tenderness

A series of three experiments were conducted to evaluate the performance of the MIRINZ (Meat Industry Research Institute of New Zealand) tenderness probe. In the first experiment, assessments on raw samples from two different muscles (longissimus lumborum and semimembranosus) were obtained using the tension head. Results indicated D50 values (torque at 50° of rotation) were the most highly related to both sensory (taste, olfactory, and feeling factors measured by a trained sensory panel) and consumer (hedonic ratings provided by untrained consumers) properties and were the values of choice, since all probe values were essentially measuring the same characteristics. In the second experiment, assessments were made using both the tension and shear heads on raw and cooked striploin (longissimus lumborum) steaks. Results confirmed probe values were essentially measuring the same characteristics, but indicated raw tension head, area 2 values (area before 50° of rotation) were more highly related to both sensory and consumer properties than were D50 values. However, probe values from raw samples did not account for a sufficient amount of variation in either sensory or consumer properties to be useful predictors of these traits, and stepwise, linear regression did not improve the predictive value of raw probe values, since all probe values were interrelated. Probe values on cooked samples indicated the shear head and area 3 values (area under the entire trace) were the most related to both sensory and consumer properties and were the values of choice. Based upon the amount of variation accounted for in sensory and consumer traits, probe area 3 values using the shear head on cooked samples provided greater predictive value for both sensory and consumer traits than Warner–Bratzler shear values obtained from steaks cooked either fresh or after being frozen and thawed. In experiment 3, correlations with individual, parametric character notes from the texture profiles of two different muscles provided no clear indication as to the textural properties being assessed by the probe. However, the MIRINZ probe appears to be a quick viable alternative to the Warner–Bratzler shear for providing an objective measure of cooked meat tenderness.     

Profile of biochemical traits influencing tenderness of muscles from the beef round

The objective of this study was to define the biochemical differences that govern tenderness and palatability of economically important muscles from the beef round using cuts with known tenderness differences. At 24h postmortem, the longissimus dorsi (LD), gracillus (GR), adductor (AD), semimembranosus (SM), sartorius (SAR), vastus lateralis (VL), and vastus intermedius (VI) muscles were removed from ten market weight beef cattle. Sensory and biochemical characteristics were determined in each cut and compared with the LD. The GR, SAR and VI had sensory traits similar to the LD while the SM, AD and VL differed. The GR, SAR, AD, and SM all had multiple biochemical characteristics similar to the LD, while the VI and AD had numerous biochemical differences. While no one biochemical characteristic can be used to predict tenderness across all muscles, analysis of the biochemical characteristics revealed that in most beef round cuts postmortem proteolysis provided a good indication of the tenderization occurring during aging.     

Establishing tenderness thresholds of Venezuelan beef steaks using consumer and trained sensory panels

The tenderness thresholds for ribeye steaks at 2 d postmortem were found by using sensory trained (n = 734 steaks) and consumers panel data (55 ribeye steaks and 296 consumers). Also, consumers’ perceptions of sensory attributes of beef steaks were assessed to establish their relationships and to determine acceptability levels. The tenderness threshold obtained by regression analysis from trained panelists was 37.98 N (3.87 kg). Customer satisfaction of 81% was obtained when beef steaks had Warner-Bratzler shear force values (WBSF) less than 40.13 N (4.09 kg), suggesting that 40.13 N could be considered as a tenderness threshold to segregate tough from tender steaks, for Venezuelan beef consumers. According to correlation and regression analyses, consumers focused on flavor to assign their overall liking score. Results suggest that consumers are able to detect changes in tenderness of steaks with different WBSF and that flavor drives customer satisfaction when the beef tenderness is acceptable.     

Correlation of consumer assessment of longissimus dorsi beef palatability with image colour,marbling and surface texture features

A new study was conducted to apply computer vision methods successfully developed using trained sensory panel palatability data to new samples with consumer panel palatability data. The computer vision methodology utilized the traditional approach of using beef muscle colour, marbling and surface texture as palatability indicators. These features were linked to corresponding consumer panel palatability data with the traditional approach of partial least squares regression (PLSR). Best subsets were selected by genetic algorithms. Results indicate that accurate modelling of likeability with regression models was possible (r² = 0.86). Modelling of other important palatability attributes proved encouraging (tenderness r² = 0.76, juiciness r² = 0.69, flavour r² = 0.78). Therefore, the current study provides a basis for further expanding computer vision methodology to correlate with consumer panel palatability data.     

Alternative pre-rigor foreshank positioning can improve beef shoulder muscle tenderness

Thirty beef carcasses were harvested and the foreshank of each side was independently positioned (cranial, natural, parallel, or caudal) 1 h post-mortem to determine the effect of foreshank angle at rigor mortis on the sarcomere length and tenderness of six beef shoulder muscles. The infraspinatus (IS), pectoralis profundus (PP), serratus ventralis (SV), supraspinatus (SS), teres major (TM) and triceps brachii (TB) were excised 48 h post-mortem for Warner–Bratzler shear force (WBSF) and sarcomere length evaluations. All muscles except the SS had altered (P < 0.05) sarcomere lengths between positions; the cranial position resulted in the longest sarcomeres for the SV and TB muscles whilst the natural position had longer sarcomeres for the PP and TM muscles. The SV from the cranial position had lower (P < 0.05) shear than the caudal position and TB from the natural position had lower (P < 0.05) shear than the parallel or caudal positions. Sarcomere length was moderately correlated (r = − 0.63; P < 0.01) to shear force.     

Relationship between commercially available DNA analysis and phenotypic observations on beef quality and tenderness

Warner–Bratzler shear force values from 560 mixed breed heifers and steers were used to determine estimates of genetic selection. Cattle were marketed from 2008 to 2011, and included five feedlot based research projects at the North Dakota State University-Carrington Research Extension Center. Samples were collected for IGENITY® analysis providing information that included selection indices and estimated breeding values for carcass traits. DNA-based test results were compared with actual carcass measurements. Marbling accounted for over 10% of the variation in WBSF while hot carcass weight was the second most influential carcass trait accounting for 4% (P < 0.01). Regression coefficients of IGENITY® molecular breeding value on phenotype for WBSF, marbling, ribeye area, yield grade, and fat thickness were low (R² = 0.14, 0.02, 0.03, 0.03, and 0.02, respectively). Therefore selecting cattle for a higher degree of marbling and feeding a diet that meets or exceeds recommended nutrients for growth are the most important factors influencing beef tenderness and acceptability.     

Meat color stability affected by barley variety fed in finishing diet to beef steers

Angus crossbred steers were assigned randomly to one of four finishing diets based on corn, Chinook, Logan, or H3 barley. Steers were harvested and after a 72-h chill, carcass quality and yield grade data were collected. Beef ribs were removed from 72 carcasses for further analysis. Ribs were aged in vacuum bag at 2 °C for 14 days. After aging three adjacent steaks (3.18 cm) were removed to determine color stability, tenderness, proximate analysis and pH. Diets fed to steers had no effect on quality and yield grade or tenderness of beef steaks. Steaks from steers fed Logan barley variety were significantly less red at 10 days of storage (Hunter a* = 24.06) than steaks from steers fed the other barley varieties (Chinook a* = 26.4; H3 a* = 28.05) or corn (a* = 26.14). Identification of barley varieties that affect color stability could result in designing diets specifically for improved color and increase the use of barley as a finishing grain.     

Relationship between beef consumer tenderness perception and Warner-Bratzler shear force

The aim of this study was to determine consumer ability to discern different levels of beef tenderness established by Warner-Bratzler shear (WBs). A panel of 220 people evaluated 60 samples of longissimus thoracis using a 5-point intensity scale (1: very tough; 5: very tender). Samples differed for commercial category, breed of animals and ageing length of meat. Shear force was measured by Instron equipped with a Warner-Bratzler device on 1.27cm diameter cores. Correlation coefficient of WBs measurements with tenderness sensory ratings was -0.72. WBs value corresponding to class 3 of the sensory tenderness was 47.77N. From this value, the range of WBs (22.96-72.59N) was split into five categories to which connect the five classes of sensory tenderness. The results suggest consumers' difficulty in discriminating category 1 (WBs>62.59N) from category 2 (WBs: 52.78-62.59N) and a greater inclination to distinguish category 5 (WBs<32.96N). As WBs category boundaries were probably too restrictive for the panel's selective ability, WBs scale was reduced to 3 categories by joining the two extreme categories (i.e. category 2 with 1 and category 4 with 5). In this case, 55.6% of consumers significantly discriminated tough from intermediate and tender meat and 62.3% distinguished tender from intermediate and tough meat (P<0.01). Hence, WBs values >52.68N and <42.87N allow classification of tough and tender beef in a sufficiently reliable way.     

Sorting for beef tenderness using high performance liquid chromatography and capillary electrophoresis: A research note

This study utilized two sampling methods to examine changes in sarcoplasmic proteins during aging of beef and their relation to tenderness. Water-soluble proteins either obtained by manually expressing exudates from the meat (drip) or by an extraction procedure using homogenization and centrifugation (ext) were analyzed for longissimus lumborum muscle using HPLC and capillary electrophoresis (CE) on days 2, 7, 10 and 14 postmortem. A peak that consistently increased with aging was identified using HPLC. Among nine peaks detected in the CE analysis, peak 9 (100kDa) that increased and peak 4 (30kDa) that decreased with aging were correlated (P<0.05) to tenderness as determined by Warner-Bratzler shear force (WBSF). For pooled data of all aging periods, drip sample explained the most variability (49%) in shear force compared to ext sample (25%) using HPLC analyses. At 2 days postmortem, a multiple linear regression model explained 83% of the variation in WBSF using CE-ext or HPLC-drip samples. Sixty percent of the variability in shear force was explained by CE-ext samples for day 7 data. The variability in shear force as explained by either drip or ext sample was less than 51 percent for 10 and 14 days postmortem data. The drip samples were comparable to ext samples in predicting WBSF values for both tough (>46N WBSF on day 2) and tender (<46N WBSF on day 2) strip loins using CE and HPLC procedure. Results suggest that a simple drip sampling may have a potential for use with either HPLC or CE analyses on day 2 postmortem for sorting carcasses into tenderness groups.     

Cooked yields, color, tenderness, and sensory traits of beef roasts cooked in an oven with steam generation versus a commercial convection oven to different endpoint temperatures

A CVap steam generation oven was compared with a Blodgett convection oven to examine effects on yields, cooked color, tenderness, and sensory traits of beef Longissimus lumborum (LL), Deep pectoralis (DP), and Biceps femoris (BF) muscles cooked to 1 of 3 temperatures (65.6, 71.1, or 76.7°C). Four roasts were cooked in the CVap for a constant time, and 2 roasts were cooked in the Blodgett until they reached target temperatures (3 replications). Cooking yields were higher (P<0.05) for BF and LL roasts in the CVap. Slice shear force (SSF) for BF roasts was lowest (P<0.05) in the CVap but lowest (P<0.05) for DP roasts in the Blodgett. No oven effect (P>0.05) was found for LL roasts. Sensory tenderness for BF roasts in the CVap was higher (P<0.05) than those in the Blodgett. Juiciness was higher (P<0.05) for LL roasts in the Blodgett. The CVap oven offers some tenderization (BF) and cooking yield advantages (BF and DP) over forced-air convection cooking.     

Imaging optical diffuse reflectance in beef muscles for tenderness prediction

The objective of this study was to investigate the potential of a novel optical reflectance imaging method to predict beef tenderness. Two-dimensional (2D) optical reflectance in beef muscles induced by a point incident light was acquired. A set of five parameters were extracted from each reflectance image to describe quantitatively the reflectance profiles. Two parameters, q and B, were derived by numerically fitting the equi-intensity contours of the reflectance pattern. Two spatial gradients were calculated along the directions parallel and perpendicular to muscle fibers and total scattering intensity was obtained by excluding the specular reflectance. This method was applied to analyze 2D images of optical diffuse reflectance in 336 beef samples obtained from 14 steers in which large variations in tenderness were generated by altering animal genetics, suspension method and aging time as well as utilizing muscles varying in their inherent tenderness. Tenderness was evaluated using Warner–Bratzler shear force (WBSF). The effects of animal breed, muscle, types of suspension, and aging were investigated and results indicate that the scattering intensity measured at 1-d was correlated (R² = 0.50 at λ = 720 nm) with 10-d WBSF in M. longissimus dorsi muscles; and the q parameters measured at 1-d was correlated (R² = 0.46 at λ = 720 nm) with 10-d WBSF in M. psoas major muscles. These results show analyzing 2D reflectance images of meat surfaces provides valuable information regarding the physical characteristics of meat that are responsible for beef tenderness.     

Comparison of various wavelet texture features to predict beef palatability

The wavelet transform can be used to characterise the surface texture of beef images in a more efficient manner than classical algorithms such as co-occurrence and run lengths. Features extracted from wavelet decompositions have been used to develop predictive models of important palatability attributes. A variety of common wavelet transforms were considered (biorthogonal, reverse biorthogonal, discrete Meyer, Daubechie, symmetric modified Daubechie and Coifman modified Daubechie) to search for the most useful texture features. A classic run length and co-occurrence algorithm was used for comparison. Using the same data analysis methods for each wavelet type, predictive models of beef acceptability, tenderness, juiciness, flavour and hardness were developed. Genetic algorithms succeeded in finding more accurate models than stepwise and manual elimination except for hardness. An accurate model of flavour (r² = 0.84) was computed. A good model of overall acceptability (r² = 0.79) was computed that fell just short of an important benchmark of accuracy. An encouraging model of juiciness (r² = 0.71) was computed showing that with additional palatability information juiciness might be accurately modelled. Tenderness proved difficult to model with only the classic model satisfying stability criteria and a poorer result (r² = 0.64) meaning substantial additional palatability information is required for accurate modelling. Hardness was particularly difficult to model. The biorthogonal wavelet produced the best model for three palatability measurements but the symmetric modified Daubechie wavelet produced the best model of overall acceptability and thus must be viewed as the most useful wavelet type.     

Influence of animal temperament and stress responsiveness on the carcass quality and beef tenderness of feedlot cattle

Three groups of steers (A, B, C) were used to study the effect of temperament (Calm, Intermediate, and Excitable) on meat quality. Temperament was based on exit velocity, pen scores, and chute scores. Temperament traits were consistent across evaluations, and values decreased (P<0.05) in magnitude over time. Increasing excitability was associated with higher (P<0.05) serum cortisol concentrations. Carcasses from cattle with calm temperaments had higher (P<0.05) 0.5h postmortem pH values than those from Intermediate and Excitable cattle (0.1 and 0.2 units, respectively). Group C Excitable steers had higher (P<0.05) WBS values than the calmer Group C steers. This trend was observed in Group A steers, although the values were not statistically different. Correlations were highest between temperament values and tenderness after 21d. Temperament influences tenderness, although the mechanism is not clear.     

Effect of age and cut on tenderness of South African beef

The tenderness characteristics of 15 primal cuts of beef of three different age groups were assessed, and the most reliable cut to predict carcass tenderness was determined. Fifteen wholesale cuts from each age group, representing the full variation in fatness, were aged, cooked and underwent sensory evaluation, shear force resistance and proximate analysis. Collagen content and solubility was determined. Percentage fat was used as a covariant during statistical analyses. Tenderness, residue and collagen solubility of all cuts decreased significantly with animal age. Collagen solubility was the largest discriminant between the three age groups, while animal age had no significant effect on collagen content. Tenderness of primal cuts from the same carcass varied considerably, with collagen content and shear force resistance as the largest discriminants between the cuts. Cuts most representative of total carcass tenderness were M. vastus lateralis, M. semimembranosus, M. gluteobiceps, M. semitendinosus and M. triceps brachii caput longum.     

Factors influencing the variation in tenderness of seven major beef muscles from three Angus and Brahman breed crosses

Beef carcasses (n=30) from 3/4 Angus (A)×1/4 Brahman (B), 1/4A×3/4B, and 1/2A×1/2B F(1) crosses were used to evaluate breed type, electrical stimulation, and postmortem aging on the M. semimembranosus (SM), M. semitendinosus (ST), M. biceps femoris (BF), M. vastus lateralis (VL), M. gluteus medius (GM), M. longissimus dorsi lumborum (LD), and M. triceps brachii (TB). Shear force values decreased with increased postmortem aging to a greater extent in steaks from 3/4A×1/4B than steaks from the other breed types. Shear force values for steaks from the round (SM, ST, BF, VL) were higher than steaks from the loin (LD, GM) and chuck (TB) for both electrically stimulated and non-electrically stimulated muscles. In the LD muscle, calpastatin activities were similar among breed types. Muscle type played the greatest role in determining tenderness.     

Prediction of beef eating quality from colour, marbling and wavelet texture features

Beef longissimus dorsi colour, marbling fat and surface texture are long established properties that are used in some countries by expert graders to classify beef carcasses, with subjective and inconsistent decision. As a computer vision system can deliver objective and consistent decisions rapidly and is capable of handling a greater variety of image features, attempts have been made to develop computerised predictions of eating quality based on these and other properties but have failed to adequately model the variation in eating quality. Therefore, in this study, examination of the ribeye at high magnification and consideration of a broad range of colour and marbling fat features was used to attempt to provide better information on beef eating quality. Wavelets were used to describe the image texture of the beef surface at high magnification rather than classical methods such as run lengths, difference histograms and co-occurrence matrices. Sensory panel and Instron analyses were performed on duplicate steaks to measure the quality of the beef. Using the classical statistical method of partial least squares regression (PLSR) it was possible to model a very high proportion of the variation in eating quality (r² = 0.88 for sensory overall acceptability and r² = 0.85 for 7-day WBS). Addition of non-linear texture terms to the models gave some improvements.     

The influence of ante-mortem treatment on relationship between pH and tenderness of beef

Longissimus lumborum et thoracis muscles from 150 cattle comprising Czech Pied breed bulls and cows were used to investigate the relationship between pH value and beef tenderness as influenced by housing before slaughter and by the hanging mode of carcasses. The pH value was measured in the loin at the end of the slaughter-line (pH(0)) and after cooling of carcasses (pH(24),pH(48)). The texture of raw meat was measured 72h post-mortem by the Warner-Bratzler shear method; the shear force was related to pH. Curvilinear relationships with coefficients of correlation r=0.64 (P<0.05) for bulls individually housed and r=0.83 for bulls housed in groups between the shear force and the pH value 24h p.m. were found with maximal shear force (the most tough meat) in the range of pH(24) from 6.2 to 6.4. Lower correlations (r=0.72 and r=0.54) between pH(24) and meat tenderness were found in the case of cows individually housed and housed in groups, respectively. No significant relationships were found between texture and pH measured at the end of the slaughter-line (r=0.59 for bulls and r=0.65 for cows) and 48h post-mortem (r=0.64 for bulls and r=0.50 for cows). The mode of housing of cattle before slaughter influenced the properties of beef significantly. Meat of bulls housed in groups before slaughter was less tender (the shear force was higher) than meat of bulls individually housed. It could be explained by higher depletion of glycogen before slaughter and thus insufficient pH decline. Generally, the meat of cows was more tender than of bulls irrespective of housing. No significant differences were found between the left and right halves of carcasses. Thus the mode of hanging had no influence on beef tenderness.     

Lateral and longitudinal characterization of color stability, instrumental tenderness, and sensory characteristics in the beef semimembranosus

The objectives of this study were to evaluate longitudinal and lateral variations in color stability and sensory characteristics of the semimembranosus from each muscle location (dorsal = D, medial = M, ventral = V), and within each steak subsequently divided into four regions (caudal-distal = CaD, cranial-distal = CrD, caudal-proximal = CaP, cranial-proximal = CrP). Instrumental color assessment indicated that L^* (lightness) values tended to be lower for the CaD and CrD regions of all muscle locations, with higher L^* values recorded for the CrP regions of muscle sections. Instrumental tenderness measurements for Meullenet–Owens razor and Warner–Bratzler shear force values were generally lower for the CaD region of the D section, whereas regions within the V section were tougher and required greater shear force. Sensory panelists perceived lower amounts of connective tissue in the CaD of the D section, which tended to be more tender, whereas higher levels of connective tissue and greater toughness occurred in the CaP and CrP regions of the M and V sections. These results provide a detailed muscle profile of the quality variation within the semimembranosus, and may be used to further enhance the fabrication and marketability of this beef subprimal cut.