Penggabungan Fitur Local Binary Patterns untuk Identifikasi Citra Tumbuhan Obat

Abstract

Identification plants automatically still be problem in obtaining a robust features. Local Binary Patterns (LBP) is an excellent descriptor for texture feature due to efficiency and robustness, but limited in the size of sampling point. In this research we propose fusion of LBP features, which incorporates additional information without sacrificing the robustnes of LBP features. Fusion of LBP features was done by two ways. In the first way, we perform a straightforward fusion by calculating histogram of multiple LBP features separately, then concatenating the multiple histograms together. In the first way we combine information provided by multiple operators by varying the size of sampling points and radius. In the second way, each histogram of LBP features are classified, and the feature fusion can be accomplished by classifier combination. Both ways, we employ probabilistic neural network (PNN) to classify LBP features. The experiment performed on medicinal plants and house plants. The fusion of LBP features are shown to be very powerful tools for improving accuracy.