Vehicle Detection using Upper Local Ternary Features with SVM Classification

Abstract

Local Ternary Pattern (LTP) is an extension of Local Binary Pattern (LBP), which is known as a standard textural descriptor for various recognition systems. There are a lot of usages of LBP and LTP in the variety of pattern recognition in different fields. It is also efficient features for vehicle detection, recognition and monitoring. This paper proposed an algorithm that applied the operator to upper LTP (upper-LTP) matrix as feature vector to calculate instead of pattern stream generation. This paper analyzes the accuracy rate between the Complemented- Uniform Local Binary Pattern (Complemented-ULBP) which is our previous work, Uniform Local Binary Pattern (ULBP), simple Local Binary Pattern (LBP) and the proposed system, Upper-LTP, with SVM classification.