Layout Analysis of Handwritten Historical Documents for Searching the Archive of the Cabinet of the Dutch Queen

Abstract

In this paper, we describe the structure and the performance of a layout analysis system developed for processing the handwritten documents contained in a large historical collection of very high importance in the Netherlands. We introduce a method based on contour tracing that generates curvilinear separation paths between text lines in order to preserve the ascenders and descenders. Our methods are relevant to research on digitization and retrieval of handwritten historical documents.

Deep neural networks for record counting in historical handwritten documents

“…Extracting information from historical handwritten documents coming from census, birth records, and other public or private record collections, is essential to reconstruct genealogies and to perform demographic researches [1] [2] [3] [4].…”

Page Segmentation of Structured Documents Using 2D Stochastic Context-Free Grammars

“…This research is focused on this second type of historical handwritten documents, concretely in marriage license books (see Figure 1). Similar structured documents have been studied in other researches [4].…”

“…In case of historical handwritten documents, the challenge in page segmentation is to detect homogeneous handwritten zones [4]. Correct detection of homogeneous textual zones is important for tackling further problems like transcription or word spotting.…”

Structure detection and segmentation of documents using 2D stochastic context-free grammars

“…In case of historical handwritten documents, the challenge in image segmentation is to detect homogeneous handwritten zones [5,6]. Correct detection of textual zones is important for tackling subsequent problems like line detection and extraction [7] and later transcription or word spotting [8].…”

“…Segmentation of this kind of documents has been approached in the past with geometrical techniques. In [5] projection profiles were mainly used for page layout analysis of documents with very satisfactory results. But for many other documents, page segmentation techniques that rely on explicit isolation of elements like characters, words or lines are often not useful.…”