Enhancement and feature extraction for images of incised and ink texts

Enhancement of historical printed document images by combining Total Variation regularization and Non-local Means filtering

“…Sattar and Tay [32] deblur noisy document images using fuzzy logic and multi-resolution approaches. Pan et al [29] correct uneven illumination and remove wood grain and shading on images of text incised on wood tablets by filtering: this allows handwritten strokes to be more easily extracted. Removing shades in document margins produced while scanning thick documents with a region growing method has been studied by Fan et al [17].…”

Reading the Readers: Modelling Complex Humanities Processes to Build Cognitive Systems

“…Novel imaging techniques have been developed at the Department of Engineering Science, University of Oxford, to analyse these texts Schenk, 2001;Molton et al, 2003;Brady et al, 2004;Pan et al, 2004). However, whilst a scrutiny of the document surface using image processing techniques provides new information, it does not provide a means of sorting through this information to provide the most likely interpretation of the data.…”

“…4 Techniques borrowed from the field of Knowledge Elicitation were used to gather quantitative and qualitative information about how papyrologists work, resulting in an in-depth understanding of the ways 3 For a discussion regarding how the language contained in the stylus tablets relates to that of the ink tablets, see Terras (2002). 4 Aalto (1945), Youtie (1963), Youtie (1968), and Bowman and Tomlin (2004) are the only discussions found (as yet) which try to describe what the papyrology process actually consists of, with some higher level discussion available in Turner (1968). Fig.…”

“…As shown in the example transcript of stylus tablet 836, there are often characters that are marked by the papyrologists as being uncertain: textual analysis of the corpus of 26,000 characters from the Vindolanda ink texts indicates that 9.9% of the characters in this corpus are highlighted as being uncertain (Terras, 2002). Bowman and Tomlin (2004) provide examples of readings which have changed dramatically between different versions of published texts.…”

A mask-based enhancement method for historical documents

“…Pan et a1. 7 correct uneven illumination and remove wood grain and shading on images of text incised on wood tablets by filtering allowing handwritten strokes to be more easily extracted. Removing shades in document margins produced while scanning thick documents with a region growing method has been studied by Fan et a1.…”