• Journal of the Korea Society of Computer and Information
• /
• v.26 no.8
• /
• pp.47-54
• /
• 2021

Given a document, keyphrase extraction is to automatically extract words or phrases which topically represent the content of the document. In unsupervised keyphrase extraction approaches, candidate words or phrases are first extracted from the input document, and scores are calculated for keyphrase candidates, and final keyphrases are selected based on the scores. Regarding the computation of the scores of candidates in unsupervised keyphrase extraction, this study proposes a method of adjusting the scores of keyphrase candidates according to the types of keyphrase candidates: word-type or phrase-type. For this, type-token ratios of word-type and phrase-type candidates as well as information content of high-frequency word-type and phrase-type candidates are collected from the input document, and those values are employed in adjusting the scores of keyphrase candidates. In experiments using four keyphrase extraction evaluation datasets which were constructed for full-text articles in English, the proposed method performed better than a baseline method and comparison methods in three datasets.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.8C
• /
• pp.1159-1174
• /
• 2004

In this paper, we present an efficient algorithm for region analysis of business card images acquired in a PDA by using DCT and information pixel density. The proposed method consists of three parts: region segmentation, information region classification, and text region classification. In the region segmentation, an input business card image is partitioned into 8 f8 blocks and the blocks are classified into information and background blocks using the normalized DCT energy in their low frequency bands. The input image is then segmented into information and background regions by region labeling on the classified blocks. In the information region classification, each information region is classified into picture region or text region by using a ratio of the DCT energy of horizontal and vertical edge components to that in low frequency band and a density of information pixels, that are black pixels in its binarized region. In the text region classification, each text region is classified into large character region or small character region by using the density of information pixels and an averaged horizontal and vertical run-lengths of information pixels. Experimental results show that the proposed method yields good performance of region segmentation, information region classification, and text region classification for test images of several types of business cards acquired by a PDA under various surrounding conditions. In addition, the error rates of the proposed region segmentation are about 2.2-10.1% lower than those of the conventional region segmentation methods. It is also shown that the error rates of the proposed information region classification is about 1.7% lower than that of the conventional information region classification method.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.29B no.10
• /
• pp.100-112
• /
• 1992

In this paper, an integrated document input/output system is developed which constructs the graphic document from a text file, converts the document into encoded facsimile data, and also recognizes printed/handwritten alphanumerics and Korean characters in a facsimile or graphic document. For an output system, we develop the method which generates bit-map patterns from the document consisting of the KSC5601 and ASCII codes. The binary graphic image, if necessary, is encoded by the G3 coding scheme for facsimile transmission. For a user friendly input system for documents consisting of alphanumerics and Korean characters obtained from a facsimile or scanner, we propose a document recognition algirithm utilizing several special features(partial projection, cross point, and distance features) and the membership function of the fuzzy set theory. In summary, we develop an integrated document input/output system and its performance is demonstrated via computer simulation.

• Journal of Information Processing Systems
• /
• v.19 no.6
• /
• pp.830-841
• /
• 2023

Tokenization is the process of segmenting the input text into smaller units of text, and it is a preprocessing task that is mainly performed to improve the efficiency of the machine learning process. Various tokenization methods have been proposed for application in the field of natural language processing, but studies have primarily focused on efficiently segmenting text. Few studies have been conducted on the Korean language to explore what tokenization methods are suitable for document classification task. In this paper, an exploratory study was performed to find the most suitable tokenization method to improve the performance of a representative spatio-temporal document classifier in Korean. For the experiment, a convolutional neural network model was used, and for the final performance comparison, tasks were selected for document classification where performance largely depends on the tokenization method. As a tokenization method for comparative experiments, commonly used Jamo, Character, and Word units were adopted. As a result of the experiment, it was confirmed that the tokenization of word units showed excellent performance in the case of representative spatio-temporal document classification task where the semantic embedding ability of the token itself is important.

• The KIPS Transactions:PartC
• /
• v.17C no.1
• /
• pp.27-36
• /
• 2010

In Web registration pages and online polling applications, CAPTCHA(Completely Automated Public Turing Test To Tell Computers and Human Apart) is used for distinguishing human users from automated programs. Text-based CAPTCHAs have been widely used in many popular Web sites in which distorted text is used. However, because the advanced optical character recognition techniques can recognize the distorted texts, the reliability becomes low. Image-based CAPTCHAs have been proposed to improve the reliability of the text-based CAPTCHAs. However, these systems also are known as having some drawbacks. First, some image-based CAPTCHA systems with small number of image files in their image dictionary is not so reliable since attacker can recognize images by repeated executions of machine learning programs. Second, users may feel uncomfortable since they have to try CAPTCHA tests repeatedly when they fail to input a correct keyword. Third, some image-base CAPTCHAs require high communication cost since they should send several image files for one CAPTCHA. To solve these problems of image-based CAPTCHA, this paper proposes a new CAPTCHA based on both image and text. In this system, an image and keywords are integrated into one CAPTCHA image to give user a hint for the answer keyword. The proposed CAPTCHA can help users to input easily the answer keyword with the hint in the fused image. Also, the proposed system can reduce the communication costs since it uses only a fused image file for one CAPTCHA. To improve the reliability of the image-text fusion CAPTCHA, we also propose a dynamic building method of large image dictionary from gathering huge amount of images from theinternet with filtering phase for preserving the correctness of CAPTCHA images. In this paper, we proved that the proposed image-text fusion CAPTCHA provides users more convenience and high reliability than the image-based CAPTCHA through experiments.

• ETRI Journal
• /
• v.39 no.4
• /
• pp.455-466
• /
• 2017

This study focuses on a method for sequential data augmentation in order to alleviate data sparseness problems. Specifically, we present corpus expansion techniques for enhancing the coverage of a language model. Recent recurrent neural network studies show that a seq2seq model can be applied for addressing language generation issues; it has the ability to generate new sentences from given input sentences. We present a method of corpus expansion using a sentence-chain based seq2seq model. For training the seq2seq model, sentence chains are used as triples. The first two sentences in a triple are used for the encoder of the seq2seq model, while the last sentence becomes a target sequence for the decoder. Using only internal resources, evaluation results show an improvement of approximately 7.6% relative perplexity over a baseline language model of Korean text. Additionally, from a comparison with a previous study, the sentence chain approach reduces the size of the training data by 38.4% while generating 1.4-times the number of n-grams with superior performance for English text.

• Journal of the Korea Society for Simulation
• /
• v.11 no.3
• /
• pp.35-48
• /
• 2002

In the work presented here, we describe a method to extract TrueType features for supporting letter recognition. Even if variously existing document processing techniques have been challenged, almost few methods are capable of recognize a letter associated with its TrueType features supporting OCR free, which boost up fast processing time for image text retrieval. By reviewing the mechanism generating digital fonts and birth of TrueType, we realize that each TrueType is drawn by its contour of the glyph table. Hence, we are capable of deriving the segment with density for a letter with a specific TrueType, defined by the number of occurrence over a segment width. A certain number of occurrence appears frequently often due to the fixed segment width. We utilize letter recognition by comparing TrueType feature library of a letter with that from input word images. Experiments have been carried out to justify robustness of the proposed method showing acceptable results.

• Korean Journal of Cognitive Science
• /
• v.13 no.3
• /
• pp.33-46
• /
• 2002

Recently, unstructured random data such as website logs, texts and tables etc, have been flooding in the internet. Among these unstructured data there are potentially very useful data such as bulletin boards and e-mails that are used for customer services and the output from search engines. Various text mining tools have been introduced to deal with those data. But most of them lack accuracy compared to traditional data mining tools that deal with structured data. Hence, it has been sought to find a way to apply data mining techniques to these text data. In this paper, we propose a text mining system which can incooperate existing data mining methods. We use text mining as a preprocessing tool to generate formatted data to be used as input to the data mining system. The output of the data mining system is used as feedback data to the text mining to guide further categorization. This feedback cycle can enhance the performance of the text mining in terms of accuracy. We apply this method to categorize web sites containing adult contents as well as illegal contents. The result shows improvements in categorization performance for previously ambiguous data.

• Journal of KIISE:Software and Applications
• /
• v.31 no.4
• /
• pp.420-438
• /
• 2004

Artificially or naturally contained texts in the natural images have significant and detailed information about the scenes. If we develop a method that can extract and recognize those texts in real-time, the method can be applied to many important applications. In this paper, we suggest a new method that extracts the text areas in the natural images using the low-level image features of color continuity. gray-level variation and color valiance and that verifies the extracted candidate regions by using the high-level text feature such as stroke. And the two level features are combined hierarchically. The color continuity is used since most of the characters in the same text lesion have the same color, and the gray-level variation is used since the text strokes are distinctive in their gray-values to the background. Also, the color variance is used since the text strokes are distinctive in their gray-values to the background, and this value is more sensitive than the gray-level variations. The text level stroke features are extracted using a multi-resolution wavelet transforms on the local image areas and the feature vectors are input to a SVM(Support Vector Machine) classifier for the verification. We have tested the proposed method using various kinds of the natural images and have confirmed that the extraction rates are very high even in complex background images.

• Journal of the Ergonomics Society of Korea
• /
• v.33 no.5
• /
• pp.355-363
• /
• 2014

Objective:The purpose of this study is to design a three-set type (Sebulsik) keyboard that is to input Korean text with no shifted keys and also compatible with the standard Korean computer keyboard or ANSI keyboard. Background: The KS computer keyboard is two-set type (Dubulsik). Existing and proposed designs of three-set type of past studies are not compatible with KS or ANSI keyboard and are complex with many redundant letters. Method: The number of Korean letters for 3-set type is analyzed. Then Korean letters are arranged with normality and with spatial compatibility to the KS Korean keyboard, and symbols were arranged to same positions with ANSI keyboard. Results: Initial consonants of 14 numbers and 6 vowels are arranged as exactly same positions of KS keyboard, and other vowels are arranged with spatial compatibility. Symbols are arranged to the same positions with ANSI keyboard, and 10 digits are confirmed and has compatibility to International standard. Conclusion: A 38-key model, 38K, is designed to require minimal keys to input Korean text with no shifted keys, increased the compatibility to the KS Korean computer keyboard. Application: Using the proposed 38-key model, 38K, it can be taken into account for keyboards in industrial production. It is applicable to user group of 3-set type Korean keyboard with more easy than past keyboards.