Dominant Local Tetra Patterns: A New Feature Descriptor for Face Recognition
A. Usha Ruby · International Journal of Emerging Trends in Engineering Research · 2020
Local binary pattern (LBP) and local ternary pattern (LTP) of the gray-level difference of the connection between the identified pixel and its nearby neighbours are encoded based on vertical and horizontal directions using the n th order local derivatives.In our proposed approach Dominant local tetra patterns (LTrPs) the n th order LTrP is computed using n th order horizontal and vertical local derivatives.The efficiency and the effectiveness of our proposed algorithm for face recognition is calculated and analysed by combining it with the Difference of Gaussian.In order to find the strength of the descriptor various images under different lighting, facial expression and pose are tested.The experimental results show that the proposed approach provides a good representation and advanced of face recognition rates than LBP and its by-products.