基于图像纹理特征的牛肉嫩度预测方法研究

在经过图像预处理,背最长肌与大理石花纹的分割,并实现大理石花纹特征值的提取后,利用灰度共生矩阵提取4个对嫩度剪切力贡献较大的纹理特征参数,并统计这些参数应用多元线性回归建立牛肉嫩度剪切力预测模型。结果表明：可见光下利用纹理特征预测牛肉嫩度的方法能够以96%的准确率实现嫩度剪切力等级的预测,具有较高的商用开发价值。     

Classification of tough and tender beef by image texture analysis

Texture features of fresh-beef images were extracted and used to classify steaks into tough and tender groups in terms of cooked-beef tenderness. Crossbred steers varying in quality were processed in a commercial plant and two short loin steaks were sampled from each carcass. One sample was used for imaging and the other was broiled for sensory evaluation of tenderness by a trained panel. The samples were segregated into tough and tender groups according to the sensory scores. A wavelet-based decomposition method was used to extract texture features of fresh-beef images. The texture feature data for 90 sample images were used to train and test sample calssifiers in a rotational leave-one-out scheme. A correct classification rate of 83.3% was obtained in cross validations. While texture features alone may not be sufficient to segregate beef products into many levels of tenderness, they can be significant members in a set of indicators that will lead to adequate tenderness prediction.     

Muscle Mineral Concentrations as Predictors of Taste Panel Sensory Attributes of Beef

M. longissimus concentrations of Zn, Co, Se, Cd, Mn, Na, Fe, Ca and Mg were used to predict taste panel sensory attributes of 144 beef rib roasts. The best equations included 4, 6, 2 and 5 predictors for flavor, juiciness, tenderness and chew test, respectively, and explained from 4.9 (tenderness) to 26% (flavor) of the variation. When all predictors were used, they explained 28.2, 25.6, 7.5 and 25.4% of the variation in flavor, juiciness, tenderness and chew test. Thus, muscle mineral concentration is as poor a predictor of sensory attributes as marbling score. If variation in taste panel sensory attributes of beef from young cattle is important, other predictors must be identified.     

Beef Marbling and Color Score Determination by Image Processing

Sixty steaks with various degrees of marbling and color were subjected to sensory evaluation and image processing. Marbling and color scores were assigned to each steak with USDA marbling score cards and a lean color guide. Images were recorded for each steak under the same conditions as used for sensory analysis. Steak images were processed for color and marbling characteristics. Image processing effectively predicted the lean color (R²= 0.86) and marbling scores (R²= 0.84). Image processing was an effective tool for determining USDA quality attributes of fresh meat.     

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.     

基于不变矩和改进SVM的牛肉大理石纹评级

针对牛肉大理石纹人工评级过程中人为误差干扰大的问题,研究利用图像处理技术提高牛肉大理石纹评级的客观性并增强自动化程度,提出基于不变矩、灰度共生矩阵和混沌蜂群优化混合核函数支持向量机(Support Vector Machine,SVM)的牛肉大理石纹评级法。首先计算牛肉大理石纹图像的不变矩和灰度共生矩阵统计量并由此构造特征向量;然后将训练和测试样本输入到混合核函数SVM,采用混沌蜂群算法优化SVM中的惩罚因子及核参数,使其分类识别性能达到最优;最后输入待评级样本进行分类识别,得到最优评级结果。大量实验结果表明:根据我国NY/T676-2010牛肉大理石纹标准图谱,评级正确率分别达到100%(一级)、93.3%(二级)、93.3%(三级)、96.7%(四级)、100%(五级)。与基于灰度矩和SVM法、基于灰度共生矩阵和BP(Back Propagation)神经网络法相比,本文所得评级正确率最高,且与专业评级师的实际评级情况最相符。     

CONSUMER EVALUATION OF THE PALATABILITY OF USDA SELECT,USDA CHOICE AND CERTIFIED ANGUS BEEF STRIP LOIN STEAKS FROM RETAIL MARKETS IN LUBBOCK,TEXAS, U.S.A.

ABSTRACT

Select, Choice and Certified Angus Beef (CAB) strip loin steaks were sold in Lubbock, TX to determine consumer acceptability. A home survey was attached to each package of steaks (return = 191 steak evaluations). Consumers did not detect differences in tenderness or flavor, but rated CAB steaks juicier (P < 0.05) than Select or Choice. However, 68% of CAB consumers rated steaks as extremely or very tender, but less than half rated Choice and Select steaks the same way. Consumer tenderness and flavor scores for Select steaks were more variable, but consumer satisfaction and tenderness acceptability did not differ (P > 0.05) between grades. Trained sensory panelists rated CAB steaks higher (P < 0.05) than Choice or Select for all palatability traits except flavor intensity. Shear force decreased (P < 0.05) by quality grade as CAB steaks sheared with the least resistance. These results showed marbling influenced objective palatability measures and decreased the variation in both objective and consumer measures of palatability.

PRACTICAL APPLICATIONS

Quality grades were established to predict beef palatability, but marbling can be an unreliable predictor of tenderness. Branded beef programs, such as Certified Angus Beef (CAB), have been developed to distinguish higher palatability products and provide more consistent products for consumers. Consumers can distinguish between levels beef tenderness and are willing to pay a premium for tender beef. Moreover, consumers find product consistency is moderately important when purchasing fresh beef. Therefore, the purpose of this study was to determine consumer acceptability based on in‐home evaluations of strip loin steaks from United States Department of Agriculture Select and Choice grades and CAB purchased from a retail supermarket environment.

     

Evaluation of pork color by using computer vision

The objective of this study was to determine the potential of computer vision technology for evaluating fresh pork loin color. Software was developed to segment pork loin images into background, muscle and fat. Color image features were then extracted from segmented images. Features used in this study included mean and standard deviation of red, green, and blue bands of the segmented muscle area. Sensory scores were obtained for the color characteristics of the lean meat from a trained panel using a 5-point color scale. The scores were based on visual perception and ranged from 1 to 5. Both statistical and neural network models were employed to predict the color scores by using the image features as inputs. The statistical model used partial least squares technique to derive latent variables. The latent variables were subsequently used in a multiple linear regression. The neural network used a back-propagation learning algorithm. Correlation coefficients between predicted and original sensory scores were 0.75 and 0.52 for neural network and statistical models, respectively. Prediction error was the difference between average sensory score and the predicted color score. An error of 0.6 or lower was considered negligible from a practical viewpoint. For 93.2% of the 44 pork loin samples, prediction error was lower than 0.6 in neural network modeling. In addition, 84.1% of the samples gave an error lower than 0.6 in the statistical predictions. Results of this study showed that an image processing system in conjunction with a neural network is an effective tool for evaluating fresh pork color.     

The relationship of physiological maturity and marbling judged according to Korean grading system to meat quality traits of Hanwoo beef females

Fifty seven carcasses from Hanwoo beef females were randomly selected by official meat graders and were sorted into three levels of maturity and marbling. Carcass data was collected for back fat thickness, longissimus area, carcass weight, meat colour, fat colour, marbling score, yield and quality grades. Mature carcasses had more yellow fat, coarser texture, a larger longissimus muscle area and lower quality grades and marbling scores (P<0.05). Carcasses with a higher marbling score had thicker fat and a higher quality grade. Carcasses with low marbling had a higher yield grade and a coarser texture (P<0.05). Higher marbling scores corresponded with lower cook and drip loss values for longissimus steaks. As the maturity of carcass was increased, the redness and lightness of meat and the yellowness of fat all tended to increase. Tenderness, flavour and overall acceptability scores for the older maturity group were lower than for younger and intermediate groups. Marbling was significantly (P<0.01) correlated with quality grade, crude fat content, cook and drip losses, and Warner-Bratzler shear force. The maturity level was also significantly (P<0.01) correlated with quality grade, fat colour, texture score, number of calves produced and milk teeth, meat redness and yellowness, fat yellowness, and Warner-Bratzler shear force. Results indicate that a low marbling group and older maturity group based on Korean grading system could negatively influence carcass traits and beef qualities of Hanwoo beef female.     

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.     

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.     

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.     

Genetic and management factors affecting beef quality in grazing Hereford steers

Attributes contributing to differences in beef quality of 206 Hereford steers finished on pasture were assessed. Beef quality traits evaluated were: Warner-Bratzler meat tenderness and muscle and fat color at one and seven days after slaughter and trained sensory panel traits (tenderness, juiciness, flavor, and marbling) at seven days. Molecular markers were CAPN1 316 and an SNP in exon 2 on the leptin gene (E2FB). Average daily live weight gain, ultrasound monthly backfat thickness gain and rib-eye area gain were estimated. Molecular markers effects on meat quality traits were analyzed by mixed models. Association of meat quality with post weaning growth traits was analyzed by canonical correlations. Muscle color and marbling were affected by CAPN1 316 and E2FB and Warner-Bratzler meat tenderness by the former. The results confirm that marker assisted selection for tenderness is advisable only when beef aging is a common practice. The most important sources of variation in tenderness and color of meat remained unaccounted for.     

Genetic and management factors affecting beef quality in grazing Hereford steers

Attributes contributing to differences in beef quality of 206 Hereford steers finished on pasture were assessed. Beef quality traits evaluated were: Warner–Bratzler meat tenderness and muscle and fat color at one and seven days after slaughter and trained sensory panel traits (tenderness, juiciness, flavor, and marbling) at seven days. Molecular markers were CAPN1 316 and an SNP in exon 2 on the leptin gene (E2FB). Average daily live weight gain, ultrasound monthly backfat thickness gain and rib-eye area gain were estimated. Molecular markers effects on meat quality traits were analyzed by mixed models. Association of meat quality with post weaning growth traits was analyzed by canonical correlations. Muscle color and marbling were affected by CAPN1 316 and E2FB and Warner–Bratzler meat tenderness by the former. The results confirm that marker assisted selection for tenderness is advisable only when beef aging is a common practice. The most important sources of variation in tenderness and color of meat remained unaccounted for.     

Changes in Cooking Losses and Sensory Attributes of Angus and Holstein Beef with Increasing Carcass Weight, Marbling Score or Longissimus Ether Extract

Cooking loss and sensory attribute changes were quantified for increases in carcass weight, marbling score and intramuscular fat in rib roasts from 74 small framed Angus and 71 Holstein steers slaughtered over a wide weight range. Cooking losses increased (P < 0.005) 2.6 percentage units for each 100 kg increase in carcass weight. Marbling score explained less than 1.2% of the variation in tenderness and was positively related (P < 0.01) to flavor of roasts from Angus but not Holsteins. Carcass weight, marbling score and intramuscular fat were more related to pan juice and total losses and explained little of the variation in sensory attributes. Therefore, alternatives to traditional indicators are needed to explain variation in sensory attributes of beef from young cattle.     

Comparison of the predictive power of beef surface wavelet texture features at high and low magnification

Beef longissimus dorsi surface texture is an indicator used in predicting beef palatability by expert graders. Computer vision systems have previously used imaging at normal view to develop surface texture features with some success. Good models of beef overall acceptability using imaging at high magnification have been recently developed. As a comparison the same surface texture features were computed from the corresponding images at normal view and used to model overall acceptability. Both sets of texture features were also combined with muscle colour and marbling features and used to model overall acceptability. Models using texture features alone were more successful at normal modality. However colour and marbling features combined much better with texture features at high modality to yield the most accurate model of overall acceptability (r² = 0.93). Accurate Partial Least Squares Regression (PLSR) models were computed at both modalities with and without inclusion of colour and marbling features. Addition of squared terms to the models failed to improve accuracy.     

Image analysis study of the perimysial connective network,and its relationship with tenderness and composition of bovine meat

Image processing method was developed to predict beef tenderness, collagen and lipids contents. The study was carried out on the semimembranosus muscle (SM). Images of SM slices were acquired under visible and ultraviolet lighting. In this work statistical technique was implemented as a method to relate the distribution of intramuscular connective tissue (IMCT), characterized by image analysis, to sensory tenderness evaluated by a trained panel and collagen and total lipids contents assessed chemically. Using Multiple Linear Regression (MLR) combining visible and ultraviolet lighting, IMCT image parameters were found to be good predictors of beef tenderness (R² = 0.89), collagen and lipids contents (respectively R² = 0.82 and R² = 0.91).     

Variation in Flavor and Textural Descriptions of Cooked Steaks from Bovine m. longissimus thoracis et lumborum from Different Production and Aging Regimes

ABSTRACT: Free choice profiling (FCP) and generalized Procrustes analysis (GPA) were used to assess flavor differences in cooked bovine m. longissimus thoracis et lumborum (LTL) steaks sourced wholesale or from steers fed the same diet. Wholesale steaks were from LTL muscles that were either unaged or aged for 93 d with high or low visible marbling from old and young grass- and grain-finished cattle. The GPA of this sensory data showed that the 1st dimension contrasted grass- and grain-finished beef and accounted for 44.1% of the total variance, whereas the 2nd dimension contrasted unaged and aged beef and accounted for 22.9% of the variance. GPA was also used to assess flavor variation in cooked LTL steaks aged 4 or 14 d postmortem from 10 steers finished on the same diet, and the 1st 2 dimensions accounted for 39% of the total variance. Univariate analysis of beef aroma, flavor intensity, tenderness, and juiciness showed only tenderness increased with aging ( P = 0.041) and Duncan's multiple comparison test on all FCP data indicated that aging beef up to 14 d shifted beef sensory characteristics toward increased tenderness, juiciness, and fatty flavor. The results of these studies showed that distinct flavors exist in beef from cattle finished in different production systems and that there is less variation in the flavor of beef from steers produced on the same farm than in beef available for general retail.     

不同烹饪方式的牛里脊质构品质比较

分析不同烹饪方式对黄牛牛里脊质构参数、脂肪酸含量的影响。购买新鲜黄牛牛里脊,分为5等份,分别做生牛里脊、炖制、烤制、油煎、真空低温烹饪。对比烹饪方式对黄牛牛里脊持水力、pH、粗蛋白、粗脂肪、感官品质、色泽、质构参数、脂肪酸含量的影响。真空低温牛里脊的持水力、pH值、粗蛋白、粗脂肪分别为89.98%、6.12、23.69 mg/g、23.69%。牛里脊色泽、嫩度、多汁性、风味评分分别为8.86、8.49、8.84、8.62分,色泽参数L^*值、a^*值、b^*值、c^*值、h0值分别为45.15、4.12、14.82、15.39、15.39;质构参数硬度、粘性、弹性、咀嚼性、内聚力、剪切力分别为8.36 N、0.62 N·s、6.98 mm、42.68 N·s、0.42 N/cm2、25.39 N;饱和脂肪酸含量、不饱和脂肪酸含量均较高。真空低温牛里脊与炖制牛里脊、烤制牛里脊、油煎牛里脊持水力、pH、粗蛋白、粗脂肪、感官品质、色泽、质构参数、脂肪酸含量相比,具有统计学差异(p<0.05)。由上述结果可知,不同烹饪的牛里脊与生牛里脊相比,存在一定的微观结构、质构变化和营养物质改变,真空低温烹饪方式下黄牛牛里脊微观结构较为完整,对质构的影响较小,更好的保留脂肪酸等营养成分,符合对黄牛牛里脊的烹饪要求。     

BOVINE MUSCLE TENDERNESS AS RELATED TO PROTEIN SOLUBILITY

ABSTRACT— The longissimus (modest degree of marbling) from forty beef ribs selected 48–56 hr post-mortem was used in two trials. Trial I involved A, C and E maturity ribs (10 each classification). Each rib was subjectively scored for texture (fresh) and adjacent longissimus samples were removed for the determination of protein solubility (fresh) and tenderness. Tenderness (cooked muscle) was measured with a Warner-Bratzler shear and taste panel. Protein solubilities were determined using 0.154M Krebs-Ringer-Bicarbonate buffer, 0.2M KCl + 0.01M K phosphate buffer, 1.1M Kl + 0.1M K phosphate buffer, and 0.03M K phosphate buffer. Trial II involved 10 A maturity ribs. The 0.2M KCl, 1.1M Kl and 0.03M K phosphate buffers as described for trial I were used for protein extraction. Additionally, sarcomere length was measured in formalin. Multiple regression equations were developed to predict tenderness in trial II. Protein solubilities were not significantly different between the carcass maturity groups although there were trends toward increased solubility as maturity increased. Tenderness tended to decrease from A to E maturity indicating a negative relationship between protein solubility and tenderness. Several significant negative correlations between protein solubility and tenderness were found in trial I (A maturity group) and trial II. Additionally, several significant negative correlations between texture and solubility were calculated. Correlations within the C and E maturity groups were variable and showed no definite trends. Multiple regression analyses showed that a combination of protein solubilities, texture score and sarcomere length accounted for 88% of the variation in shear force and 72% of the variation in taste panel tenderness.