Temporal and longitudinal variances in wood log cross-section image analysis

In this work two practical issues of biometric log recognition using log end images are investigated: Temporal and longitudinal variances of log cross-sections (CSs). These variances are related to the requirement of robustness for biometric characteristics. A texture feature-based fingerprint matching technique is used to compute matching scores between CS images. Our experimental evaluation is based on the temporal and longitudinal variances of 35 slices of a single tree log which where captured at four time delayed sessions. Results indicate, that biometric log recognition using log end images is robust and is able to overcome both issues. This work contributes to the development of a biometric log recognition system by showing that a texture feature-based matching technique is applicable to log CSs.