Recognition of large-set printed Hangul (Korean script) by two-stage backpropagation neural classifier

Cho, Sung Bae; Kim, Jin Hee

doi:10.1016/0031-3203(92)90147-b

Cited by 15 publications

(6 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In recent years, neural networks have shown a high recognition accuracy for digital [9], natural scene [10]- [12], and handwritten characters [13], [14]. Regarding Korean Hangul, methods of modeling its structural features have been proposed [15], [16] and recognition by neural networks have been used [17].…”

Section: Related Workmentioning

confidence: 99%

On the Ability of a CNN to Realize Image-to-Image Language Conversion

Baba

Uchida

Iwana

2019

2019 International Conference on Document Analysis and Recognition (ICDAR)

View full text Add to dashboard Cite

The purpose of this paper is to reveal the ability that Convolutional Neural Networks (CNN) have on the novel task of image-to-image language conversion. We propose a new network to tackle this task by converting images of Korean Hangul characters directly into images of the phonetic Latin character equivalent. The conversion rules between Hangul and the phonetic symbols are not explicitly provided. The results of the proposed network show that it is possible to perform image-to-image language conversion. Moreover, it shows that it can grasp the structural features of Hangul even from limited learning data. In addition, it introduces a new network to use when the input and output have significantly different features.

show abstract

Section: Related Workmentioning

confidence: 99%

On the Ability of a CNN to Realize Image-to-Image Language Conversion

Baba

Uchida

Iwana

2019

2019 International Conference on Document Analysis and Recognition (ICDAR)

View full text Add to dashboard Cite

show abstract

“…consonants on their left (top). The shape of a vowel varies depending on its position and accompanying symbols [6].…”

Section: Table 1 Classification Of Korean Phonemesmentioning

confidence: 99%

“…Then by a graph search algorithm, DP matching costs are calculated between segment 5 and each of segments 2, 4, 3, 6, 7, 8, and 9. The final sequence (5,4,6,8) minimum cost is extracted as a set of segments corresponding to the line segment ' '. To apply DP matching to 2D patterns we extend this algorithm by concatenating DP matching cost planes according to the branch and bound control strategy.…”

Section: Branch and Bound Algorithm Based On Dp Matchingmentioning

confidence: 99%

“…Although this algorithm can be constructed and employed for recognition of alphanumeric characters, it is difficult to di-rectly apply the algorithm to recognition of Korean characters in which more than 11,000 characters consisting of different types of phonemes are possible. The structural complexity and diverse font variations of Korean characters have been topics of research [4,5], and various recognition techniques of Korean characters have been proposed in recent decades [6][7][8][9][10]. The handwritten Korean character recognition method robust to distortion was proposed by using adaptive nonlinear shape matching [11].…”

Section: Introductionmentioning

confidence: 99%

“…The handwritten Korean character recognition method robust to distortion was proposed by using adaptive nonlinear shape matching [11]. Also algorithms based on neural network [6,7] have been proposed; however, these algorithms were focused on recognition of Korean characters, not on the analysis of a document. Also, statistical algorithms were presented to analyze documents consisting of characters and graphic regions [9,10], in which recognition of printed characters and some constrained handwritten characters was investigated.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations