Bleed-Through Removal from Degraded Documents Using a Color Decorrelation Method

Abstract

Abstract.A color decorrelation strategy to improve the human or automatic readability of degraded documents is presented. The particular degradation that is considered here is bleed-through, that is, a pattern that interferes with the text to be read due to seeping of ink from the reverse side of the document. A simplified linear model for this degradation is introduced to permit the application of very fast decorrelation techniques to the RGB components of the color data images, and to compare this strategy to the independent component analysis approach. Some examples from an extensive experimentation with real ancient documents are described, and the possibility to further improve the restoration performance by using hyperspectral/multispectral data is envisaged.

Manuscript Bleed-through Removal via Hysteresis Thresholding

“…We have included the three blind bleed-through removal methods mentioned earlier plus a few additional approaches: Otsu and Sauvola [11] are examples of document binarization methods, with the former being global and the latter adaptive, while ICA [14] represents a general statistical technique. For our method, Hyst otsu uses Otsu's threshold as t h , while Hyst best uses a slightly more corrosive threshold that achieves better bleed-through removal.…”

“…Concentrating on blind methods, some of the more prominent efforts include directional wavelets by Wang et al [16], symmetric orthogonalization by Tonazzini et al [14], and recursive unsupervised classification by Fadoua et al [6]. The first is based on the difference in slant between ink and bleed-through strokes.…”

Restoring Ink Bleed-Through Degraded Document Images Using a Recursive Unsupervised Classification Technique

“…All the techniques cited above treat a particular case of degraded document image, where foreground and interfering strokes characters are oriented differently, which is not always the case. Other more flexible techniques exist, among which, we can cite techniques based on Independent Component Analysis [8], adaptive binarization [9], self-organizing maps [10], color analysis [11].…”

Pairwise Markov Model Applied to Unsupervised Image Separation

“…In this section we consider the problem of show-through and bleed-through effects in scanned document [10,15,11,14,16]. In this situation, we propose to apply our algorithm which allows a good separation automatically without any manual intervention.…”

“…Independently, Pairwise Markov based models have proved their efficiency for recovering hidden discrete data, with application to image segmentation [4,9]. Inspired from the problem formulation of recovery of scanned documents subject to show-through and bleed-through effects [10,15,11,14,16], we make in this paper a link between the two approaches. We also highlight in this context the interest of Pairwise Markov Chain (PMC) models, which have been recently introduced in [4], and we study the performance of the corresponding methods.…”