Prediction Method of Beef Marbling Standard Number Using Parameters Obtained from Image Analysis for Beef Ribeye
Prediction Method of Beef Marbling Standard Number Using Parameters Obtained from Image Analysis for Beef Ribeye
复制标题
利用图像分析获得的肋眼牛肉参数预测牛肉大理石花纹标准数的方法
DOI：
10.2508/chikusan.70.107
复制
发表时间：
1999
期刊：
影响因子：
--
通讯作者：
S. Miyoshi
中科院分区：
文献类型：
--
作者：
K. Kuchida;S. Tsuruta;L. D. Vleck;Mitsuyoshi Suzuki;S. Miyoshi
研究了影响牛肉大理石花纹标准值与图像分析大理石花纹标准值差异的因素。本文采用106头日本黑阉牛的眼肌面积照片，观察了牛的眼肌面积。用图像分析法计算眼肌区域大理石纹率、大理石纹颗粒面积和形状评分的平均值和标准差，以及眼肌分为4、9、25和100个分区的小区域大理石纹率的标准差。结果表明：眼肌区域大理石纹率为0.01、0.05、0.1、0.5和1.0cm ~ 2。以图像分析性状和眼肌面积的25个独立协变量为自变量，采用逐步回归法，得到了以BMSUB和BMSFAT之间的差异为因变量的多元回归方程。方程中使用的独立协变量的最终数量限制为3个。BMSFAT与BMSUB之间的差值范围为-3 ~+4，在±1以内的百分比为67.0%，而BMSUB与BMS数量之间的差值范围为-2 ~+2，在±1以内的百分比为91.5%。这些结果表明，通过不仅使用脂肪面积的比率，而且还使用其他图像分析特征，预测BMS数量的准确性有所提高。
Factors affecting the difference between the Beef Marbling Standard (BMS) number assigned by examiners (BMSSUB) and the BMS number estimated from marbling percentage by image analysis (BMSFAT) were investigated. Pictures of ribeye area of 106 Japanese Black steers with BMSSUB were used. Marbling percentage in ribeye area, means and standard deviations of the area and of the form score for marbling particles classified into 5 levels (over 0.01, 0.05, 0.1, 0.5, and 1.0cm2), and standard deviations of marbling percentages in small areas which were obtained by dividing the ribeye into 4, 9, 25, and 100 partitions were calculated by image analysis. Multiple regression equations with the difference between BMSSUB and BMSFAT as the dependent variable were obtained by a stepwise method starting with 25 independent covariates for image analysis traits and ribeye area. The final number of independent covariates used in the equation was limited to three. The range of the difference between BMSFAT and BMSSUB was from -3 to +4 and the percentage of the differences within±1 was 67.0%, while the range of the difference between BMSSUB and the BMS number which was calculated from a multiple regression equation was from -2 to +2 and percentage of the differences within±1 was 91.5%, These results show that the accuracy of prediction for BMS number has improved by using not only the ratio of fat area but also other image analysis traits.