VAST: International Symposium on Virtual Reality, Archaeology and Intelligent Cultural Heritage - Short and Project Papers

Detection and Classification of Petroglyphs in Gigapixel Images -- Preliminary Results

Markus Seidl and Christian Breiteneder

DOI: 10.2312/PE/VAST/VAST11S/045-048

Abstract:
With the advances of digital photography, the number of high quality images of rock panels containing petroglyphs grows steadily. Different time-consuming manual methods to determine and document the exact shapes and spatial locations of petroglyphs on a panel have been carried out over decades. We aim at automated methods to a) segment rock images with petroglyphs, b) classify the petroglyphs and c) retrieve similar petroglyphs from different archives. In this short paper, we present an approach for the unsolved problem of rock art image segmentation. A first evaluation demonstrates promising results.

Categories and Subject Descriptors (according to ACM CCS): I.4.6 [Image Processing and Computer Vision]: Segmentation-Pixel Classification

full Paper

first Page