Fusion of multiple handwritten word recognition techniques
Introduction
Many successful techniques have been developed to recognize well-segmented and isolated handwritten characters and numerals. Excellent recognition results (Lee, 1995, Avi-Itzhak and Diep, 1995, Lee, 1996, Cho, 1997, Gilloux, 1993) have been achieved; however, their success has not carried onto the handwritten word recognition domain (Gader et al., 1995, Blumenstein and Verma, 1999, Gader et al., 1996, Suen et al., 1993, Srihari, 1993, Bozinovic and Srihari, 1989, Yanikoglu and Sandon, 1993, Chiang, 1998). This has been ascribed to the difficult nature of unconstrained handwritten words, including the diversity of character patterns, ambiguity and illegibility of characters, and the overlapping nature of many characters in a word (Blumenstein and Verma, 1999, Gader et al., 1996).
Researchers have used different feature extraction, segmentation and classification algorithms (Gader et al., 1995, Blumenstein and Verma, 1999, Gader et al., 1996, Casey and Lecolinet, 1996, Strathy et al., 1993, Martin et al., 1993, Eastwood et al., 1997, Lu and Shridhar, 1996, Otsu, 1979, Han and Sethi, 1995, Yanikoglu and Sandon, 1998) to achieve better recognition rates for handwritten words. The results obtained by different techniques vary significantly because many complex procedures such as preprocessing, thinning, slant correction, segmentation and classification are required to recognize unconstrained handwriting. A technique that uses very strict preprocessing and removes noise may recognize some words but it may fail to recognize words that have lost information discarded by thinning, slant correction or segmentation. On the other hand, a technique without strict preprocessing or a better segmentation algorithm may recognize those words that were not recognized by the previous technique. Therefore, various techniques in conjunction with conventional and intelligent algorithms make different errors and produce different recognition results. It is very interesting that, even if they produce similar results, the mistakes made by them might be different.
Fusion is one of the powerful methods for improving recognition rates produced by various techniques. It takes advantage of different errors produced by different techniques, emphasizes the strengths and avoids weaknesses of individual techniques. Researchers have found (Gader et al., 1996) that in many real-world applications, it is better to fuse multiple techniques to improve results.
This paper proposes a modified Borda count (MBC) to fuse three techniques developed at two different institutes using different segmentation and neural network algorithms. Experimental results on the Centre of Excellence in Document Analysis and Recognition (CEDAR) database from the individual and combined techniques are provided. A comparison of results with conventional Borda (Gader et al., 1996), majority rule (Verikas et al., 1999), averaging (Verikas et al., 1999) and the Choquet integral (Gader et al., 1996) is also included.
The remainder of the paper is broken down into five sections. Section 2 describes the proposed technique, Section 3 provides experimental results, a discussion of the results takes place in Section 4 and a conclusion is drawn in Section 5.
Section snippets
Proposed technique for fusion
This section describes the proposed approach to combine three handwritten word recognition techniques (MUMLP, GUMLP, MURBF) using a modified Borda count based on ranks and confidence values. An overview of the technique is provided in Fig. 1.
Experimental results
The experiments were conducted on cursive handwritten words taken from the CEDAR benchmark database (Hull, 1994). The database is easily available from CEDAR on one CD-ROM. The database contains real-world zip codes, city and state names from handwritten postal envelops. It was obtained from United States Postal Services (USPS). To make comparison easier with other researchers, the database is divided into training and test words. The training and test sets contain 3106 and 317 words,
Discussion
The results from individual techniques are presented in Table 3. As can be seen, the MUMLP achieved best word recognition results as an individual technique. The reason it achieved the best results was that the MUMLP used compatibility scores and very complicated rules to decide whether a union is valid or invalid during the dynamic programming based matching. Also it used very strict preprocessing which removed all types of noise from words and resized them to a fixed size. GUMLP and MURBF
Conclusion
Fusion of three different techniques has been presented in this paper, producing excellent results. The main contribution of this paper is a modified Borda count for fusion of multiple techniques using the different conventional and intelligent algorithms. The conventional Borda count, majority rule, averaging, Choquet integral and the proposed approach were tested and compared on handwritten words from the CEDAR benchmark database. The Borda count proposed in this paper, based on word rank and
Acknowledgements
We would like to thank J. Liu and W. Chen from the University of Missouri and M. Blumenstein from Griffith University for their help in conducting the experiments for our segmentation techniques. Also we would like to thank the University of Missouri and Griffith University for supporting this research.
References (23)
A hybrid neural model in handwritten word recognition
Neural Networks
(1998)- et al.
Fusion of handwritten word classifiers
Pattern Recognition Lett.
(1996) Research into the new generation of character and mailing address recognition systems at the French Post Office Research Center
Pattern Recognition Lett.
(1993)Multilayer cluster neural network for totally unconstrained handwritten numeral recognition
Neural Networks
(1995)- et al.
Character segmentation in handwritten words – an overview
Pattern Recognition
(1996) Recognition of handwritten and machine-printed text for postal address interpretation
Pattern Recognition Lett.
(1993)- et al.
Soft combination of neural classifiers: a comparitive study
Pattern Recognition Lett.
(1999) - et al.
Segmentation of off-line cursive handwriting using linear programming
Pattern Recognition
(1998) - et al.
High accuracy optical character recognition using neural networks with centroid dithering
IEEE Trans. Pattern Anal. Machine Intell.
(1995) - et al.
Neural-based solutions for the segmentation and recognition of difficult handwritten words from a benchmark database
Off-line cursive script word recognition
IEEE Trans. Pattern Anal. Machine Intell.
Cited by (39)
Effect of ensemble classifier composition on offline cursive character recognition
2013, Information Processing and ManagementCitation Excerpt :Diversity of handwriting styles is identified to be the main reason for the unsatisfactory recognition accuracy. This brings in a lot of research in recent times (Arica & Yarman-Vural, 2001, 2002; Camastra, 2007; Camastra & Vinciarelli, 2001; Chim, Kassim, & Ibrahim, 1998; Cruz, Cavalcanti, & Ren, 2010; Fujisawa, 2008; Kavallieratou, Sgarbas, Fakotakis, & Kokkinakis, 2003; Oliveira, Sabourin, Bortolozzi, & Suen, 2002; Verma, Gader, & Chen, 2001; Zhang, 2006) to investigate different features for cursive character recognition that are invariant to handwriting styles. A survey of different feature extraction methods is available in Arica and Yarman-Vural (2001, 2002), Fujisawa (2008).
Decision fusion of horizontal and vertical trajectories for recognition of online Farsi subwords
2013, Engineering Applications of Artificial IntelligenceCitation Excerpt :The achieved accuracy of the test and training sets were 64.1% and 77.8%, respectively, for x(t) signal. Combining classifiers is an approach to improve the recognition performance (Ho et al., 1994; Verma et al., 2001). In this work, two models were trained for each subword.
Binary segmentation algorithm for English cursive handwriting recognition
2012, Pattern RecognitionCitation Excerpt :Rule-based modules validate every over-segmentation point against closed area, average character size, contour code of left character, and density. Verma et al. [24] from also proposed over-segmentation and validation approach to solve the cursive handwriting recognition problem. Three different validation experts are employed to produce the ranks and confidence values, which are the ingredients of Borda count.
Segment confidence-based binary segmentation (SCBS) for cursive handwritten words
2011, Expert Systems with ApplicationsCitation Excerpt :Despite sleepless research in OffCHR for decades, the performance of the state-of-the-art OffCHR is below the industrial standard to accommodate the real world problems (Das & Dulger, 2009; Djioua & Plamondon, 2009; Lu & Tan, 2008; Suresh & Arumugam, 2007; Zhang, Yip, Brown, & Lim Tan, 2009). The researchers in this field agree that the main contributor of the low OffCHR performance is the segmentation (Arica & Yarman-Vural, 2001, 2002; Bar-Yosef, Mokeichev, Kedem, Dinstein, & Ehrlich, 2009; Blumenstein & Verma, 2001; Casey & Lecolinet, 1996; Vellasques, Oliveira, Britto, Koerich, & Sabourin, 2008a, 2008b; Verma, Gader, & Chen, 2001; Zhao, Chi, Shi, & Yan, 2003). Segmentation is a process to discriminate each letter from others, prior to recognition into electronic character representations.
Document Analysis and Recognition: A survey
2023, TechRxivA Survey on Applying Automated Recognition of Touchscreen Stroke Gestures to Children's Input
2020, Interacting with Computers