• International Journal of Contents
• /
• v.8 no.3
• /
• pp.83-93
• /
• 2012

In this paper, we propose a novel method that automatically generates real character images to familiarize existing OCR systems with new fonts. At first, we generate synthetic character images using a simple degradation model. The synthetic data is used to train an OCR engine, and the trained OCR is used to recognize and label real character images that are segmented from ideal document images. Since the OCR engine is unable to recognize accurately all real character images, a substring matching method is employed to fix wrongly labeled characters by comparing two strings; one is the string grouped by recognized characters in an ideal document image, and the other is the ordered string of characters which we are considering to train and recognize. Based on our method, we build a system that automatically generates 2350 most common Korean and 117 alphanumeric characters from new fonts. The ideal document images used in the system are postal envelope images with characters printed in ascending order of their codes. The proposed system achieved a labeling accuracy of 99%. Therefore, we believe that our system is effective in facilitating the generation of numerous character samples to enhance the recognition rate of existing OCR systems for fonts that have never been trained.

• KIISE Transactions on Computing Practices
• /
• v.21 no.2
• /
• pp.138-143
• /
• 2015

The Boyer-Moore algorithm is a single pattern string matching algorithm that is widely used in various applications such as computer and internet security, and bioinformatics. This algorithm is computationally demanding and requires high-performance parallel processing. In this paper, we propose a parallelization and performance optimization methodology for the BM algorithm on a GPU. Our methodology adopts an algorithmic cascading technique. This results in significant reductions in the mapping overheads for the threads participating in the parallel string matching. It also results in the efficient utilization of the multithreading capability of the GPU which improves the load balancing among threads. Our experimental results show that this approach achieves a 45-times speedup at maximum, in comparison with a serial execution.

• KIPS Transactions on Software and Data Engineering
• /
• v.2 no.2
• /
• pp.119-122
• /
• 2013

Repetitive strings such as periods have been studied vigorously in so diverse fields as data compression, computer-assisted music analysis, bioinformatics, and etc. In bioinformatics, periods are highly related to repetitive patterns in DNA sequences so called tandem repeats. In some cases, quite similar but not the same patterns are repeated and thus we need approximate string matching algorithms to study tandem repeats in DNA sequences. In this paper, we propose a new definition of approximate periods of strings based on distance sum. Given two strings $p({\mid}p{\mid}=m)$ and $x({\mid}x{\mid}=n)$, we propose an algorithm that computes the minimum approximate period distance based on distance sum. Our algorithm runs in $O(mn^2)$ time for the weighted edit distance, and runs in O(mn) time for the edit distance, and runs in O(n) time for the Hamming distance.

• Journal of information and communication convergence engineering
• /
• v.7 no.4
• /
• pp.419-430
• /
• 2009

Reading comprehension (RC) systems are to understand a given text and return answers in response to questions about the text. Many previous studies extract sentences that are the most similar to questions as answers. However, texts for RC tests are generally short and facts about an event or entity are often expressed in multiple sentences. The answers for some questions might be indirectly presented in the sentences having few overlapping words with the questions. This paper proposes a conceptual graph matching method towards RC tests to extract answer strings. The method first represents the text and questions as conceptual graphs, and then extracts subgraphs for every candidate answer concept from the text graph. All candidate answer concepts will be scored and ranked according to the matching similarity between their sub-graphs and question graph. The top one will be returned as answer seed to form a concise answer string. Since the sub-graphs for candidate answer concepts are not restricted to only covering a single sentence, our approach improved the performance of answer extraction on the Remedia test data.

• Journal of Internet Computing and Services
• /
• v.3 no.3
• /
• pp.1-9
• /
• 2002

The various researches are investigated for transforming byte code into objective machine code which can be implemented in the specific processor using classical compiling methods to improve the execution speed of the JAVA language. The code generation techniques using pattern matching can generate more high-quality code than code expansion techniques. We provide, in this research, the standardized pattern describing methods and pattern matching techniques that can be used to generate the register-based inter-language which is for the effective native code generation from byte code. And we designed and realized the inter-code transformer with which we can generate the high-quality register-based inter-code using standardized pattern described formerly.

• The Journal of the Acoustical Society of Korea
• /
• v.21 no.1
• /
• pp.82-88
• /
• 2002

In this paper, we proposed a audio indexing method that is used wavelet transform for audio data retrieval. It is difficult for audio data to make a efficient audio data index because of its own particular properties, such as requirement of large storage, real time to transfer and wide bandwidth. An audio data in del using wavelet transform make it possible to index and retrieval by using the particular wavelet transform properties. Our proposed indexing method doesn't separate data to several blocks. Therefore we use both high-pass and low-pass parts of last level coefficient of wavelet transform. Audio data indexing is made by applying the string matching algorithm to high-pass part and zero-crossing histogram to low-pass part. These are transformed to the continued strings, Through this method, we described a retrieval efficiency. The retrieval method is done by comparing the database index string to the query string and then data of minimum values is chosen to the result. Our simulation decided proper comparative coefficient and made known changing of retrieval efficiency versus audio data length. The results show that the proposed method improves retrieval efficiency compared to conventional method.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.12
• /
• pp.61-67
• /
• 2021

In this paper, we propose a method of extracting and recognizing related information for research from images of the unstructured pulmonary function test papers using character detection and recognition techniques. Also, we develop a post-processing method to reduce the character recognition error rate. The proposed structuring method uses a character detection model for the pulmonary function test paper images to detect all characters in the test paper and passes the detected character image through the character recognition model to obtain a string. The obtained string is reviewed for validity using string matching and structuring is completed. We confirm that our proposed structuring system is a more efficient and stable method than the structuring method through manual work of professionals because our system's error rate is within about 1% and the processing speed per pulmonary function test paper is within 2 seconds.

• Journal of KIISE:Software and Applications
• /
• v.37 no.10
• /
• pp.788-801
• /
• 2010

Advancing of mobile device is remarkable, so the research on mobile input device is getting more important issue. There are lots of input devices such as keypad, QWERTY keypad, touch and speech recognizer, but they are not as convenient as typical keyboard-based desktop input devices so input strings usually contain many typing errors. These input errors are not trouble with communication among person, but it has very critical problem with searching in database, such as dictionary and address book, we can not obtain correct results. Especially, Hangeul has more than 10,000 different characters because one Hangeul character is made by combination of consonants and vowels, frequency of error is higher than English. Generally, suffix tree is the most widely used data structure to deal with errors of query, but it is not enough for variety errors. In this paper, we propose fast approximate Korean word searching system, which allows variety typing errors. This system includes several algorithms for applying general approximate string searching to Hangeul. And we present profanity filters by using proposed system. This system filters over than 90% of coined profanities.

• Journal of Digital Contents Society
• /
• v.17 no.6
• /
• pp.555-563
• /
• 2016

Science and technology contents (research papers, patents, reports) are the most common reference material for researchers involved in research and development in the fields of science and technology. Based on various search elements (title, abstract, keyword, year of publication, name of journal, name of author, publisher, etc.), many services are available for users to search science and technology contents and bibliographic information owned by libraries. Authority data on organization name can be useful as an element for author identification and as an element to search for results produced by specific organizations. However, organization name is not taken into account by current search services for domestic academic information and bibliographic records. This study analyzes organization name data contained in the metadata of science and technology contents, which are the basis of the establishment of authority data, and proposes a method and system based on string containment and exact string matching.

• Journal of KIISE:Computer Systems and Theory
• /
• v.32 no.6
• /
• pp.268-278
• /
• 2005

We consider constructing the generalized suffix way of strings A and B when the suffix arrays of A and B are given, j.e., merging two suffix arrays of A and B. There are efficient algorithms to merge some special suffix arrays such as the odd array and the even array. However, for the general case that A and B are arbitrary strings, no efficient merging algorithms have been developed. Thus, one had to construct the generalized suffix arrays of A and B by constructing the suffix array of A$\#$B$\$$ from scratch, even though the suffix ways of A and B are given. In this paper, we Present efficient merging algorithms for the suffix arrays of two arbitrary strings A and B drawn from constant and integer alphabets. The experimental results show that merging two suffix ways of A and B are about 5 times faster than constructing the suffix way of A$\#$B$\$$ from scratch for constant alphabets. Our algorithms include searching all suffixes of string B in the suffix array of A. To do this, we use suffix links in suffix ways and we developed efficient algorithms for computing the suffix links. Efficient computation of suffix links is another contribution of this paper because it can be used to solve other problems occurred in bioinformatics that should search all suffixes of a given string in the suffix array of another string such as computing matching statistics, finding longest common substrings, and so on. The experimental results show that our methods for computing suffix links is about 3-4 times faster than the previous fastest methods.