Division 3: The lean meat content in a pig carcass according to the Union reference method is predicted by a formula on the basis of online variables extracted from an image made by ultrasound. More than 50 online variables are obtained from the image analysis. The statistical analysis reduces the information to two components, each of which is a linear combination of the same six online variables. The final formula is expressed by online variables: - Commission Implementing Decision of 17 July 2014 authorising methods for grading pig carcases in Sweden and repealing Decision 97/370/EC (notified under document C(2014) 4946) (Only the Swedish text is authentic) (2014/476/EU)

Y = 55,2971 - 0,27747 \times R2P4 - 0,24594 \times R2P11 + 4,59557 \times R2P12 - 0,22981 \times R2P15 + 0,11882 \times R3P5 - 0,11719 \times R4P3

where:

R2P4

p2_selected_fat_mm. The P2 fat measure at the selected position in mm.

R2P11

minpair_value. A filter mask that selects two regions 14 cm apart is applied to the cross-section vector. This is the minimum of the filter result vector.

R2P12

P2_skew. Relation of the selected P2 and the unselected P2. The actual point used is a little closer to the center, to make the value more tolerant to very tilted carcasses. Value is always greater than or equal to 1,0.

R2P15

minpair_value v2. A second version of the minpair value.