Anisotropic Diffusion and Curve Evolution for Segmentation of Color Images in Cultural Heritage Luigi Cinque 1 and Rossella Cossu 2 1 Dipartimento Informatica, Unversit´a degli Studi, Sapienza, Via Salaria 113, 00185 Roma, Italy 2 Istituto per le Applicazioni del Calcolo-CNR, Via dei Taurini 19, 00185 Roma, Italy r.cossu@iac.cnr.it http://www.iac.cnr.it Abstract. We propose an innovative procedure for extracting decay regions from color images of stony materials. The use of appropriate im- age analysis techniques can offer an important contribution to be used together with the traditional methodologies for studying and diagnos- ing the decay of stony materials that constitute ancient monuments. The presented approach is constituted by the PDE (Partial Differen- tial Equations) model of anisotropic diffusion and by the level set/fast marching method. The anisotropic diffusion is applied in order to limit the smoothing at the zones of high gradient. In the segmentation pro- cess, the numerical technique of the level set/fast marching is applied in order to extract from the image only the color region examined. The study case concerns impressive remains of the city of Aosta (Italy). Keywords: Color image segmentation, anisotropic diffusion, fast marching, level set. 1 Introduction The image analysis can represent an important tool for Cultural Heritage scien- tists to analyze and collect information about the preservation state of historical monuments. In the images of ancient buildings, pixels of different color charac- terize the decay areas [1]. On the wall facades of the monuments, in fact, we note different types of decay such as oxidation, sediment and/or cavities which are recognizable by the color feature. The acquired images, related to the facades of a monument, show, for example, reddish stains due to oxidation, whitish ones due to the sediment of saline efflorescence or generally dark/black ones due to lack of material as in cavities. For this reason a crucial step to extract decay re- gions, characterized by chromatic alteration, can be the application of an image segmentation strategy to color images of stony materials. In this paper we present the procedure developed for extracting decay zones from color images of stony materials. It is based on the PDE model anisotropic diffusion [2], [3] and on the level set/fast marching method [4],[5]. A. Petrosino (Ed.): ICIAP 2013, Part I, LNCS 8156, pp. 572–581, 2013. c  Springer-Verlag Berlin Heidelberg 2013