Eli Turkel

We deal with the general issue of handling statistical data in archaeology for the purpose of deducing sound, justified conclusions. The employment of various quantitative and statistical methods in archaeological practice has existed from its beginning as a systematic discipline in the 19th century (Drower 1995). Since this early period, the focus of archaeological research has developed and shifted several times. The last phase in this process, especially common in recent decades, is the proliferation of collaboration with various branches of the exact and natural sciences. Many new avenues of inquiry have been inaugurated, and a wealth of information has become available to archaeologists. In our view, the plethora of newly obtained data requires a careful reexamination of existing statistical approaches and a restatement of the desired focus of some archaeological investigations. We are delighted to dedicate this article to Israel Finkelstein, our teacher, adviser, colleague, and friend, who is one of the father figures of this ongoing scientific revolution in archaeology (e.g., Finkelstein and Piasetzky 2010, Finkelstein et al. 2012, 2015), and wish him many more fruitful years of research.

The problem of finding a prototype for typewritten or handwritten characters belongs to a family of "shape prior" estimation problems. In epigraphic research, such priors are derived manually, and constitute the building blocks of "paleographic tables". Suggestions for automatic solutions to the estimation problem are rare in both the Computer Vision and the OCR/Handwriting Text Recognition communities. We review some of the existing approaches, and propose a new robust scheme, suitable for the challenges of degraded historical documents. This fast and easy to implement method is employed for ancient Hebrew inscriptions dated to the First Temple period.

Our research team enjoyed the privilege of collaborating with Benjamin Sass over a period of several years. We are happy to dedicate this article to him and wish to express our gratitude for what has been both a prodigious and enjoyable experience. The purpose of our joint endeavor has been the introduction of modern techniques from computer science and physics to the realm of Iron Age epigraphy. One of the most important issues addressed during our cooperation was the topic of facsimile creation. Facsimile creation is a necessary preliminary step in the process of deciphering and analyzing ancient inscriptions. Several manual facsimile construction techniques are currently in use: drawing upon collation of the artifact; outlining on transparent paper overlaid on a photograph of the inscription; and computer-aided depiction via software such as Adobe Photoshop, Adobe Illustrator, Gimp or Inkscape (see Summary section below for software web links). Despite their importance for the field of epigraphy, little attention has thus far been devoted to the methodology of facsimile creation (though the recent comprehensive treatment by Parker and Rollston 2016). Recent decades have seen rapid development and consolidation of various computerized image processing algorithms. Among the most basic and popular tasks in this field is the creation of a black-and-white version of a given image, denoted as image binarization (see Fig.1a–b). Often, such a binarized image is used as a first step for further image processing missions, such as Optical Character Recognition (OCR), texts digitization and text analysis tasks. An algorithmic creation of binarizations can therefore be seen as another method of facsimile creation. Furthermore, a relatively new sub-domain of image processing, Historical Imaging and Processing (HIP), specializes in handling antique documents of different types, periods and origins. Accordingly, binarization algorithms stemming from HIP are even more suitable for archaeological purposes.

Chan-Vese is an important and well-established segmentation method. However, it tends to be challenging to implement, including issues such as initialization problems and establishing the values of several free parameters. The paper presents a detailed analysis of Chan-Vese framework. It establishes a relation between the Otsu binarization method and the fidelity terms of Chan-Vese energy functional, allowing for intelligent initialization of the scheme. An alternative, fast, and parameter-free morphological segmentation technique is also suggested. Our experiments indicate the soundness of the proposed algorithm.

This article discusses the quality assessment of binary images. The customary, ground truth based methodology, used in the literature is shown to be problematic due to its subjective nature. Several previously suggested alternatives are surveyed and are also found to be inadequate in certain scenarios. A new approach, quantifying the adherence of a binarization to its document image is proposed and tested using six different measures of accuracy. The measures are evaluated experimentally based on datasets from DIBCO and H-DIBCO competitions, with respect to different kinds of binarization degradations.

The discipline of First Temple Period epigraphy (the study of writing) relies heavily on manually-drawn facsimiles (black and white images) of ancient inscriptions. This practice may unintentionally mix up documentation and interpretation. The article proposes a new method for evaluating the quality of the facsimile. It is based on a measure, comparing the image of the inscription to the registered facsimile. Some empirical results, supporting the methodology, are presented. The technique is also relevant to quality evaluation of other types of facsimiles and binarization in general.

The relationship between the expansion of literacy in Judah and composition of biblical texts has attracted scholarly attention for over a century. Information on this issue can be deduced from Hebrew inscriptions from the final phase of the first Temple period. We report our investigation of 16 inscriptions from the Judahite desert fortress of Arad, dated ca. 600 BCE—the eve of Nebuchadnezzar’s destruction of Jerusalem. The inquiry is based on new methods for image processing and document analysis, as well as machine learning algorithms. These techniques enable identification of the minimal number of authors in a given group of inscriptions. Our algorithmic analysis, complemented by the textual information, reveals a minimum of six authors within the examined inscriptions. The results indicate that in this remote fort literacy had spread throughout the military hierarchy, down to the quartermaster and probably even below that rank. This implies that an educational infrastructure that could support the composition of literary texts in Judah already existed before the destruction of the first Temple. A similar level of literacy in this area is attested again only 400 y later, ca. 200 BCE.

A binarization of challenging historical inscription is improved by means of sparse methods. The approximation is based on a binary dictionary learned by k-medians and k-medoids algorithms from a clear source. Some preliminary results show superiority to the existing binarization with respect to fine features such as strokes continuity, deviations from a straight line, edge noise and the presence of stains. The k-medians dictionary-learning scheme shows a robust behavior when initial patches database is reduced.

The authors present a new method of writer identification, employing the full power of multiple experiments, which yields a statistically significant result. Each individual binarized and segmented character is represented as a histogram of 512 binary pixel patterns—3 × 3 black and white patches. In the process of comparing two given inscriptions under a "single author" assumption, the algorithm performs a Kolmogorov–Smirnov test for each letter and each patch. The resulting p-values are combined using Fisher's method, producing a single p-value. Experiments on both Modern and Ancient Hebrew data sets demonstrate the excellent performance and robustness of this approach.