Efficient Transcript Mapping to Ease the Creation of Document Image Segmentation Ground Truth with Text-Image Alignment

One of the major issues in document image processing is the efficient creation of ground truth in order to be used for training and evaluation purposes. Since a large number of tools have to be trained and evaluated in realistic circumstances, we need to have a quick and low cost way to create the corresponding ground truth. Moreover, the specific need for having the correct text correlated with the corresponding image area in text line and word level makes the process of ground truth creation a difficult, tedious and costly task. In this paper, we introduce an efficient transcript mapping technique to ease the construction of document image segmentation ground truth that includes text-image alignment. The proposed text line transcript mapping technique is based on Hough transform that is guided by the number of the text lines. Concerning the word segmentation ground truth, a gap classification technique constrained by the number of the words is used. Experimental results prove that using the proposed technique for handwritten documents, the percentage of time saved for ground truth creation and text-image alignment is more than 90%.

[1]  Ioannis Pratikakis,et al.  Text line detection in handwritten documents , 2008, Pattern Recognit..

[2]  Lambert Schomaker,et al.  Text-image alignment for historical handwritten documents , 2009, Electronic Imaging.

[3]  C. V. Jawahar,et al.  Content-level Annotation of Large Collection of Printed Document Images , 2007, Ninth International Conference on Document Analysis and Recognition (ICDAR 2007).

[4]  Horst Bunke,et al.  Automatic segmentation of the IAM off-line database for handwritten English text , 2002, Object recognition supported by user interaction for service robots.

[5]  Georgios Louloudis,et al.  ICDAR 2009 Handwriting Segmentation Contest , 2009, ICDAR.

[6]  John D. Hobby,et al.  Matching document images with ground truth , 1998, International Journal on Document Analysis and Recognition.

[7]  C. Halatsis,et al.  Line And Word Segmentation of Handwritten Documents , 2008 .

[8]  Sargur N. Srihari,et al.  Mapping Transcripts to Handwritten Text , 2006 .

[9]  Alejandro Héctor Toselli,et al.  Viterbi Based Alignment between Text Images and their Transcripts , 2007, LaTeCH@ACL 2007.

[10]  James Allan,et al.  Text alignment with handwritten documents , 2004, First International Workshop on Document Image Analysis for Libraries, 2004. Proceedings..

[11]  B. Noble,et al.  On certain integrals of Lipschitz-Hankel type involving products of bessel functions , 1955, Philosophical Transactions of the Royal Society of London. Series A, Mathematical and Physical Sciences.

[12]  Venu Govindaraju,et al.  Transcript mapping for handwritten Arabic documents , 2007, Electronic Imaging.

[13]  R. Manmatha,et al.  Aligning Transcripts to Automatically Segmented Handwritten Manuscripts , 2006, Document Analysis Systems.

[14]  Basilios Gatos,et al.  Handwriting Segmentation Contest , 2007, Ninth International Conference on Document Analysis and Recognition (ICDAR 2007).

[15]  Chun-Jen Chen,et al.  A linear-time component-labeling algorithm using contour tracing technique , 2004, Comput. Vis. Image Underst..

[16]  Ioannis Pratikakis,et al.  Text line and word segmentation of handwritten documents , 2009, Pattern Recognit..

[17]  Sergios Theodoridis,et al.  Keyword-guided word spotting in historical printed documents using synthetic data and user feedback , 2007, International Journal of Document Analysis and Recognition (IJDAR).

[18]  C. V. Jawahar,et al.  Content-level Annotation of Large Collection of Printed Document Images , 2007 .

[19]  Apostolos Antonacopoulos,et al.  Semantics-based content extraction in typewritten historical documents , 2005, Eighth International Conference on Document Analysis and Recognition (ICDAR'05).

[20]  Bin Zhang,et al.  Transcript mapping for historic handwritten document images , 2002, Proceedings Eighth International Workshop on Frontiers in Handwriting Recognition.