Grouping of Crack Patterns using Proximity and Characteristic Rules

F.S. Abas and K. Martinez (UK)


Computer Vision, Pattern Analysis and Recognition, CrackAnalysis, Data Clustering


In this paper, we present a 2-stage approach to connected curve grouping. The algorithm was developed and demon strated on crack-detected images of paintings. Some fea tures are left undetected and this tends to produce dis connected curves. In order to extract high-level features for content-based application, these supposedly connected curves have to be grouped together. It is one of the many steps needed to produce a content-based platform for dig ital analysis of crack patterns in paintings particularly for classification purpose. The prime objective of the grouping algorithm is to segment or partition areas of an image to produce reliable representations of content. The first stage of the algorithm utilizes the Minimum Bounding Rectangle (MBR) of a crack network as means to decide on merging using a proximity rule. We demonstrate the use of the both the rotated and the un-rotated MBR. In the second stage, curve characteristics represented by the rotated MBR such as the dimension ratio, the axis of minimum inertia, object centroid and node density are used as features for an N dimensional clustering.

Important Links:

Go Back