Comparison of GLCM and DCT features for Malayalam Palm Leaf Character Recognition

This study presents a comprehensive comparison of two distinct feature extraction techniques, namely Gray Level Co-occurrence Matrix (GLCM) and Discrete Cosine Transform (DCT), for the purpose of Malayalam palm leaf character recognition. The study focuses on the extraction of features from palm leaf images, employing GLCM to capture spatial relationships of pixel intensities and DCT to represent the frequency domain information. Extracted features are trained using different statistical classifiers. A thorough evaluation is conducted to assess the performance of both sets of feature vectors in the context of Malayalam palm leaf character recognition. The experimental results demonstrate the efficacy of DCT than GLCM in capturing the unique characteristics of Malayalam characters, providing valuable insights for the development of robust recognition systems in the domain of historical document analysis and digitization. Random Forest (RF), AdaBoost, and Decision tree (DT) demonstrate strong performance on the DCT feature set.