Proceedings of Student/Faculty Research Day, CSIS, Pace University, May 6, 2005 A Greek Pottery Shape and School Identification and Classification System Using Image Retrieval Techniques Gulsebnem Bishop, Sung-Hyuk Cha and Charles Tappert {gf44944w, scha, ctappert}@pace.edu ABSTRACT In this paper we propose an image-based pottery shape and school identification and classification system for an unknown pottery or fragment. This system is designed to assist archaeologists and students to identify and record objects quickly and accurately. We present several image retrieval and computer vision techniques and describe their applications within the domain of archaeological studies by utilizing a large digital library of pottery photos. Twenty pottery shapes and four pottery schools are identified with shape and color-based image retrieval techniques, respectively. The system analyses and compares extracted features to determine the five closest database images, and then presents them to the user for final decision. This is the first pottery study to combine the two different techniques – shape and color-based image retrieval – to identify multiple characteristics of an unknown pottery image or a pottery fragment. The database contains about one thousand pottery photos obtained from online digital libraries. Experiments on identifying the correct pottery shape and pottery schools yielded pproximately 98% accuracy based on the whole pottery image. a Key words: image retrieval, jigsaw, pottery classification, template matching 1. INTRODUCTION The study of pottery is one of the most important and complex tasks done by archaeologists. A significant amount of information about a culture can be gleaned from studying pottery. Analyzing the physical aspects of a pot, or a fragment, can give us detailed information about the history behind the artifact. By observing the shape and the size, we can establish a date range. By observing the decorations, we can narrow the date range of a vessel. Pottery contains a great deal of information; therefore, as it is hard not to agree with Kampel & Sablatnig: it is necessary to find a new and more effective means of analysis.[1] 1.1 Current Study of Pottery Most commonly found object in an excavation, and probably the most useful, is the pottery or pottery fragments. From these artifacts it is possible to build an entire chronology examining the stylistic changes occurring in each level of habitation. They provide us with information about technical, artistic and decorative changes and developments which gives us a general understanding of the civilization as a whole. They also give us information about the people’s daily lives such as what they ate, what they produced and traded and who they worshipped. The study of pottery start from the time it is unearthed and never ends. Usually, the process goes through 8 stages; retrieval, washing, examination, registration, cutting, description, drawing and publication, but this process changes from excavation to excavation. [16] The retrieval, analysis and the classification of the pottery found in excavations are very labor and time intensive. This study has several modus operandi. First, it will serve as an educational tool for a novice archaeologist to identify and study artifacts or fragments easily and quickly. Second, it can serve as a valuable tool in excavations for identification, classification, and reconstruction of fragments. This may lead to significant archaeological discoveries. Third, thousands of pottery fragments are unearthed every year, and are usually discarded after being recorded. It is often impossible to photograph or record the measurements of these fragments, or even assign them a particular category or class. This system will provide a quick, inexpensive and objective way of documenting and classifying these fragments. Finally, it can assist in identification and analysis of pottery decorations. 1.2 Previous Research Studies and research on finding a recording and a retrieval system for archaeological finds go back to the early 70’s. Most studies in this decade concentrated on how computers and computer applications could be used in the field of archaeology. While Lengyel C2.1