Classification of Facial Expressions based on Transitions Derived from Third Order Neighborhood LBP

The present paper extended the LBP transitions derived from second-order neighbourhood on to third order neighbourhood LBP (TN-LBP) and derived transitions on Trapezoid patterns for facial expression classification. The TN-LBP forms four Trapezoid Patterns (TP) i.e. top left, bottom right and top right, bottom left. So far no researcher carried out work on classification problem based on transitions on third-order neighborhood LBP. The present paper derived transitions on the two reciprocal “Trapezoids of TN-LBP (T-TN-LBP) i.e. top left vs. bottom right. Each of these Trapezoids on TN-LBP will have five pixies and each of them will have 25 i.e 32 patterns. The present paper derived transitions on two symmetric T-TN-LBP. Based on this, facial expression recognition algorithm is built. The proposed approach is compared with the existing methods. Graphics & Vision Classification of Facial Expressions based on Transitions Derived from Third Order Neighborhood LBP Dr. Vakulabharanam Vijaya Kumar α , Gorti Satyanaraya Murty σ & Pullela S V V S R Kumar ρ AbstractThe present paper extended the LBP transitions derived from second-order neighbourhood on to third order neighbourhood LBP (TN-LBP) and derived transitions on Trapezoid patterns for facial expression classification. The TNLBP forms four Trapezoid Patterns (TP) i.e. top left, bottom right and top right, bottom left. So far no researcher carried out work on classification problem based on transitions on thirdorder neighborhood LBP. The present paper derived transitions on the two reciprocal “Trapezoids of TN-LBP (T-TNLBP) i.e. top left vs. bottom right. Each of these Trapezoids on TN-LBP will have five pixies and each of them will have 2 i.e 32 patterns. The present paper derived transitions on two symmetric T-TN-LBP. Based on this, facial expression recognition algorithm is built. The proposed approach is compared with the existing methods.