Optimizing Wood Type Identification Using Local Binary Pattern (LBP): Exploring the Impact of Neighborhood Configurations

Wood is a vital natural resource for various household products such as cabinets, wardrobes, chairs, tables, etc. Identifying wood types is crucial, as it involves recognizing fine features such as pore arrangement, shape, pore frequency, and wood radius. This identification is essential for determining the appropriate wood type for manufacturing specific furniture items, as different types of wood possess varying levels of strength and distinct patterns. The research focuses on wood types commonly used in furniture production. The first step is converting the RGB images of the wood into grayscale, followed by feature extraction, and classification's step. The Local Binary Pattern (LBP) extracted features from grayscale images before identifying the wood type with K-Nearest Neighbor (KNN). After obtaining the labels from the classification results, the performance is evaluated by measuring the accuracy. This study proposes an optimization approach for wood-type identification. The optimization process explores the impact of the neighborhood number on the performance LBP method. This study achieves the highest accuracy in 97.14% with a neighborhood of 3, 6, and 7 using K=1 in KNN classification.