Abstract— Archaeology is the scientific study of the remnants of human civilization, and involves exploring the lives of ancient peoples by examining their waste. Numerous researchers proposed numerous methods and ideas for the (semi or automated) reconstruction of a (usually large) number of broken irregular fragments. Hence, this paper aims to provide an in-depth review of the most notable publications on computer applications related to the area of classification and reconstruction of archaeological pottery fragments using two-Dimensional images during the period between the early 1970s and 2014. The considered publications were classified on the basis of the study type, which is divided into two categories: studies that focus on classification of ancient fragments into groups, and those that focus on the reconstruction of archaeological fragments. This paper reviews and analyzes the most relevant works published according to the extracted features, classification processes involved, matching techniques that have been implemented for the restoration of pottery objects to their original form, as well as the yielded results.


I. INTRODUCTION
The reconstruction of fragments of relics is important because it helps archaeologists access and make inferences about ancient civilizations [1][2]. It is considered among the difficult and unsolved problems within the field of computer vision [3]. Millions of archaeological fragments are stored in large warehouses, and remain without being assembled for many years [4][5], This is due to assemble the fragments are seem similar to assemble the pieces of a jigsaw puzzle [6][1]. However, much work has been done on finding a solution for the problem of the reconstruction of archaeological fragments with the aid of computers, where this task is implemented in a two-phase process, involving classification of the fragments into similar groups, and their reconstruction to the form of the original object [4][12][11]. Currently, the classification of the fragments into similar groups is performed manually [4]. Knowing that the artifacts are often found at archaeological excavation sites in a fractured state, and are mixed with each other randomly, the process of manual classification is a difficult and time-consuming task. Sometimes this task may take years of tedious work [5][3][7], especially if the object has many irregular fragments. The reconstruction process of the fragments into the original object involves a comparison of two fragments in terms of geometric, color and texture features [4]. Hence, it is worthy to further investigate and find solutions for classifying and reassembling fragmented archeological fragments to their original form. Therefore, numerous publications have dealt with the problem of finding a system for the recording and retrieval of archaeological fragments, which go back to the early 1970s [10]. Many researchers are challenged to this task of reassembling archaeological fragments to their original form in an accurate manner, mainly due to the high value in the information that represents past civilizations and cultures. Therefore, the aim of this paper is to review the notable studies in the body of related literature in order to enable researchers to develop and propose new algorithms to achieve better results than previous algorithms.

II. THE STRUCTURE OF PAPER
In order to review previous studies, this paper concentrates on the classification and reconstruction of 2D historical pottery fragments. We divided the previous methods into two groups according to the type of study. The first group includes the articles on classification of ancient pottery fragments into groups by computer assistance, while the second group consists of the reconstruction of archaeological fragments.