Earticle

Character segmentation is a preprocessing step in many offline handwriting recognition systems. In this paper, Chinese characters are categorized into seven different structures. In each structure, the character size with the range of variations is estimated considering typical handwritten samples. The component removal and merge criteria are presented to remove punctuation symbols or to merge small components which are part of a character. Finally, the criteria for segmenting the adjacent characters concerning each other or overlapped are proposed.

자동차 번호판 인식은 지능형 교통시스템에서 핵심적인 역할을 담당한다. 따라서 효율적으로 자동차 번호판의 숫자 및 문자영역을 검출하는 것은 매우 중요한 과정이다. 본 연구에서는 딥러닝과 의미론적 영상분할 알고리즘을 적용 하여 효과적으로 자동차 번호판의 번호영역을 검출하는 방법을 제안한다. 제안된 방법은 화소 투영과 같은 전처리과정 없이 번호판 영상에서 바로 숫자 및 문자영역을 검출하는 알고리즘이다. 번호판 영상은 도로 위에 설치된 고정 카메라로 부터 획득한 영상으로 날씨 및 조명변화 등을 모두 포함한 다양한 실제 상황에서 촬영된 것을 사용하였다. 입력 영상은 색상변화를 줄이기 위해 정규화하고 실험에 사용된 딥러닝 신경망 모델은 Vgg16, Vgg19, ResNet18 및 ResNet50이 다. 제안방법의 성능을 검토하기 위해 번호판 영상 500장으로 실험하였다. 학습을 위해 300장을 할당하였으며 테스트용 으로 200장을 사용하였다. 컴퓨터모의 실험결과 ResNet50을 사용할 때 가장 우수하였으며 95.77% 정확도를 얻었다.

License plate recognition plays a key role in intelligent transportation systems. Therefore, it is a very important process to efficiently detect the number and character areas. In this paper, we propose a method to effectively detect license plate number area by applying deep learning and semantic image segmentation algorithm. The proposed method is an algorithm that detects number and text areas directly from the license plate without preprocessing such as pixel projection. The license plate image was acquired from a fixed camera installed on the road, and was used in various real situations taking into account both weather and lighting changes. The input images was normalized to reduce the color change, and the deep learning neural networks used in the experiment were Vgg16, Vgg19, ResNet18, and ResNet50. To examine the performance of the proposed method, we experimented with 500 license plate images. 300 sheets were used for learning and 200 sheets were used for testing. As a result of computer simulation, it was the best when using ResNet50, and 95.77% accuracy was obtained.

In ubiquitous computing environments, many applications are devised to provide autonomous services. To make services autonomous, each service has to inevitably recognize the deployed devices or objects in ubiquitous computing environments. In order to provide an autonomous service based on vision system, we address an efficient character segmentation algorithm by means of introducing the locally adaptive binarization and the object labeling by Connected Components Analysis (CCA). The proposed binarization technique carries out the grayscale-to-binary image conversion with block-based processing to reflect the local variation of images. The object labeling by CCA algorithm is applied to the binarized image with the morphological operation in order to improve the object connectivity. Experimental results show that the proposed algorithm can segment each object reasonably even with atypical or erratic form. Therefore, the proposed algorithm can be useful for development of various applications in ubiquitous computing environments.

In this paper we present a novel container ISO-code recognition method which uses vertical edge information, a spatial structure window, and texture clustering. The vertical edge information is extracted using a top-hat transform. The candidate region and type of ISO-Code is obtained using a Spatial Structure Window (SSW) which wraps around the vertical edges. The ISO-Code is extracted using texture clustering by the K-Means algorithm which is then recognized by a Back-propagation Neural Network (BP). Experiments confirmed the robustness of the recognition algorithm on real images and videos.

In this paper we present a novel container ISO-code recognition method which uses vertical edge information, a spatial structure window, and texture clustering. The vertical edge information is extracted using a top-hat transform. The candidate region and type of ISO-Code is obtained using a Spatial Structure Window (SSW) which wraps around the vertical edges. The ISO-Code is extracted using texture clustering by the K-Means algorithm which is then recognized by a Back-propagation Neural Network (BP). Experiments confirmed the robustness of the recognition algorithm on real images and videos.

In this paper, the methods of correcting skew vehicle license plate and segmenting characters in plate are discussed first. An approach making use of self-organizing map (SOM) is introduced to find the tilt angle of plate which simultaneously educes seven important points with coordinates being elements of weight matrix. After necessary processing to corrected plate, a character segmentation algorithm based on the shortest distance classification is presented, which takes advantage of exactly seven points gained from SOM as class centers. In the next place, a hybrid algorithm cascading two steps of template matching is utilized to recognize Chinese characters segmented from the license plates, which is based on the connected region feature and standard deviation feature extracted from sample corresponding to each template. Experimental results show that the proposed method can be implemented efficiently.

The character segmentation technology for printed documents plays an important role in optical character recognition, ticket information identification, postal code identification, automatic license plate recognition and so on. In this paper, a Chinese characters segmentation algorithm for complicated printed documents is proposed for the application in paper watermarking system. In this application, the algorithm aims to achieve high accuracy Chinese character segmentation and high consistent segmentation between the digital version images and print-scanned version images for the same documents. In this method, three main steps are included: connected regions recognition, connected regions merging, and fine-gained segmentation. Experiments show the effectiveness of the proposed algorithm.

Hand written Character Recognition is area of research since many years. Automation of existing manual system is need of most industries as well as government areas. Recognition of hand written characters is a demand for many fields. In this paper we have discussed our approach for hand written character segmentation. This paper discusses various methodologies to segment a text based image at various levels of segmentation. This paper serves as a guide for people working on the text based image segmentation area of Computer Vision. First, the need for segmentation is justified in the context of text based information retrieval. Then, the various factors affecting the segmentation process are discussed. Followed by the levels of text segmentation are explored. Also, the available techniques with their advantages and weaknesses are reviewed, along with directions for quick referral are suggested. At last, we have given our approach to text segmentation in brief.

Due to the rapid development of mobile devices equipped with cameras, instant translation of any text seen in any context is possible. Mobile devices can serve as a translation tool by recognizing the texts presented in the captured scenes. Images captured by cameras will embed more external or unwanted effects which need not to be considered in traditional optical character recognition (OCR). In this paper, we segment a text image captured by mobile devices into individual single characters to facilitate OCR kernel processing. Before proceeding with character segmentation, text detection and text line construction need to be performed in advance. A novel character segmentation method which integrates touched character filters is employed on text images captured by cameras. In addition, periphery features are extracted from the segmented images of touched characters and fed as inputs to support vector machines to calculate the confident values. In our experiment, the accuracy rate of the proposed character segmentation system is 94.90%, which demonstrates the effectiveness of the proposed method.

A connected component analysis from a binary image is a popular character segmentation method but occasionally fails to segment the characters owing to image noise and uneven illumination. A multimethod binarization scheme that incorporates two or more binary images is a novel solution, but selection of binarization methods has never been analyzed before. This paper reveals the best combination of binarization methods and parameters and presents an in-depth analysis of the multimethod binarization scheme for better character segmentation. We carry out an extensive quantitative evaluation, which shows a significant improvement over conventional single-method binarization methods. Experiment results of six binarization methods and their combinations with different test images are presented.

Off-line handwritten character recognition is in difficulty of incomplete preprocessing because it has not dynamic and timing information besides has various handwriting, extreme overlap of the consonant and vowel and many error image of stroke. Consequently off-line handwritten character recognition needs to study about preprocessing of various methods such as binarization and thinning. This paper considers running time of watershed algorithm and the quality of resulting image as preprocessing For off-line handwritten Korean character recognition. So it proposes application of effective watershed algorithm for segmentation of character region and background region in gray level character image and segmentation function for binarization image and segmentation function for binarization by extracted watershed image. Besides it proposes thinning methods which effectively extracts skeleton through conditional test mask considering running time and quality. of skeleton, estimates efficiency of existing methods and this paper's methods as running time and quality. Watershed image conversion uses prewitt operator for gradient image conversion, extracts local minima considering 8-neighborhood pixel. And methods by using difference of mean value is used in region merging step, Converted watershed image by means of this methods separates effectively character region and background region applying to segmentation function. Average execution time on the previous method was 2.16 second and on this paper method was 1.72 second. We prove that this paper's method removed noise effectively with overlap stroke as compared with the previous method.

인쇄체 문자 인식에서 접합 문자는 주요한 에러 발생의 원인이다. 본 논문에서는 접합 문자를 분할하기 위해 두 개의 분할 비용을 정의한다. 첫째, 절단 비용은 한 패턴을 분할하는 데 얼마나 많은 블랙픽셀이 분리되어야 하는가이다. 둘째, 접선 비용은 분할선이 얼마나 많은 블랙 픽셀과 화이트 픽셀사이를 지나가는가이다. 폰트 분류기는 접합 문자의 후보 문자를 제공한다. 후보 문자의 문자 폭은 접합 문자를 분리하기 위한 기준선을 제공하며, 그 기준선 부근의 픽셀들이 분할 가능 영역을 나타낸다. 절단 비용의 최소값과 접선 비용의 최대값이 되는 지점이 최종적으로 접합 문자를 분할하는 위치이다. 이렇게 정의된 절단 비용과 접선 비용을 가지고 접합 문자를 분할하면 보다 정확한 문자 분할을 하여 문자 인식에서 에러 발생을 줄일 수 있다.

제철소에서 생산된 철판의 물류 자동화 시스템을 구축하는데 있어 각 철판에 기록된 정보를 정확하게 파악하여 효과적인 적재 및 운반 시스템을 갖추는 것이 관건이다. 이를 위하여 본 논문에서는 생산 및 운반 정보가 기록된 철판 영상을 분석하여 자동화 시스템을 구축하기 위한 문자 추출 기법을 제안한다. 철판 영상은 가로 방향으로 철판의 정보를 나타내는 문자열을 포함하고 있고, 이러한 문자열이 세로 방향으로 배치되어 있기 때문에 각 문자를 분리하기 위해서는 2차원 기법을 적용하도록 한다. 영상을 획득하는 환경에 따라 문자 분리 성능에 미치는 영향을 최소화하기 위하여 국부 특징을 반영한 효율적인 이진화 기법을 제안하고, 잡음 등에 민감한 특성을 제거하기 위하여 CCA(Connected Component Analysis)를 이용한 문자 여부를 판단하는 방법을 제안한다. 이렇게 분석된 영상에 대해 2차원 투영 기법을 적용하여 철판에 기록된 각 문자를 정확하게 분리하도록 한다. 제안된 문자 추출 기법은 높은 성능을 나타내면서도 저 복잡도를 가지도록 설계하여 제한된 자원을 이용하는 기기에 효과적으로 응용될 수 있다.

This paper proposes a character segmentation algorithm of steel plate images composed of adaptive binarization by the SCW (Sliding Concentric Windows) technique, the object labelling by CCA (Connected Component Analysis), and 2D projection method. The SCW technique carries out the grayscale-to-binary image conversion in consideration of local characteristics of images. The character decision algorithm followed by the labelling technique by CCA (Connected Component Analysis) determines the character area effectively reducing the noise effect. The 2D projection with horizontal and vertical directions produces a tight bounding box for a character based on the cross points. Experimental results indicate that the proposed algorithm segments the characters in steel plate images effectively. The proposed algorithm can be applied to the devices with limited resources due to its excellent performance and low complexity.

문자 영역의 추출은 명함 등 문서의 정형화된 문자 인식, 비전 기반 감시 시스템에서의 간판, 부호 등의 자연영상에서의 문자 인식 등 다양한 분야에 활용될 수 있다. 우리가 관심을 갖는 문자는 간판이나 이름표 등 다른 이에게 정보를 전달해주는 기능을 하는 것으로 그 전배경의 구분이 명확하다. 이러한 특징은 히스토그램의 변화와 관련이 되어 있으며 본 논문에서는 그 변화를 분석함으로써 문자 영역 추출 방법을 제안한다.

This paper presents a new method of segmenting characters in a car licence plate which is less influenced by illumination variation. It uses an anisotropic filter to reduce the lighting noise and a histogram specification scheme to obtain the binary image. Anisotropic smoothing filter process the input images, which are acquired under different lighting conditions, so that they may have similar image quality. The enhanced performance of the proposed algorithm has been proved by the experiment.

This paper describes about the printed character segmentation and recognition system for slabs in steel manufacturing process. To increase the recognition rate, it is important to improve success rate of character segmentation. Since Slabs front area surface are not uniform and surface temperature is very high, marked characters not only undergo damages but also have much noise. On the other hand, since almost marked characters are very thick and the space between characters is only about 10 $^{\sim}$ 15 mm, there are many touching characters. Therefore appropriate character image preprocessing and segmentation algorithm is needed. In this paper we propose a multi-local thresholding method for damaged character restoration, a modified touching character segmentation, algorithm for marked characters. Finally a effective Multi-Class SVM is used to recognize segmented characters.

본 연구에서는 접합 문자를 분할하는 알고리듬을 소개한다. 문자 인식기는 문자를 인식하기 위해 문자 분할을 전 처리 단계에 필요로 하는데, 문자 분할은 높은 수행력을 위해 문자 인식 결과를 필요로 한다. 이 딜레마를 해결하기 위해서는 문자 분할과 문자 인식, 이 두 문제를 동시에 해결하는 방법이 필요하다. 이를 위해 본 논문에서는 문자 분할 전에 접합 문자 내에 있는 소속 문자를 인식하고 문자를 분할하는 새로운 문자 분할 방법을 제시한다. 본 연구에서 제시한 문자 분할 알고리즘은 접합 문자 내에 있는 소속 문자를 문자 분할 전에 인식하기 위해 측면 윤곽을 정의하고, 그 히스토그램을 구해 프로토타입에 있는 단일 문자의 측면 윤곽 히스토그램과 비교 계산하여 가장 적은 거리차를 가지는 단일 문자를 분할 문자의 일차 후보로 내정하여, 분할 비용을 가지고 접합문자를 분할한다.

운송 컨테이너 영상에서 문자 영역 분할 오류는 컨테이너 자동화 처리를 위한 운송 컨테이너 식별자 인식시스템의 성능에 심각한 영향을 주는 요소 중의 하나이므로, 운송 컨테이너 식별자 인식 시스템의 성능을 향상시키기 위해서는 보다 정확한 문자 영역의 분할이 요구된다. 그러므로 본 연구에서는 문자 영역으로 오인할 수 있는 운송 컨테이너 영상에 포함된 여러 가지 불필요한 영상들을 형태학적 연산에 의하여 제거하고 식별자 영역 즉, 문자열 영역만을 분리한다. 이렇게 분리된 문자열 영역의 임계값을 구하여 이진화 한 후에 개별 문자로 분할하는 방법을 제안한다. 또한, 제안한 방법으로 실험한 결과는 컨테이너 식별자의 각 문자들이 정확하게 분할되었음을 보여준다.

Extracting the character region(container identifier) in the image of a shipping container is one of the key factors in a system for identifying a shipping container automatically To improve the performance of the automatic recognition system for identifying a shipping container, thus a method partitioning the character region more correctly and efficiently is needed. In this paper, an efficient method is proposed to extract only the character region in the image of a shipping container. The proposed method removes noises that are not possibly related to the character using morphological operation, then the image is binarized using the threshold value that is determined from the image obtained previous step. Finally individual character area is extracted from the binary image. Also experiments are conducted to verify the efficiency of the proposed method. The results show that the proposed method partitions the character region correctly from container images.

자동차 번호판 인식시스템은 영상획득, 번호판 추출, 전처리(이치화), 문자영역분할, 문자인식 등의 5가지 핵심부분으로 구성되어 있다. 따라서 자동차 번호판 인식시스템의 최종 인식률은 각 단계의 성능에 따라 직접적인 영향을 받는다. 본 논문은 컴퓨터 비젼의 한 분야인 영상처리 기법을 이용한 이치화된 자동차 번호판의 문자영역 추출에 관한 연구로서 문자 인식단계에서 높은 인식률을 확보하기 위해서 가장 중요한 입력 데이터의 상태를 보다 깨끗하게 정확하게 분리하는데 변형된 Run Length Coding 기법을 이용하여 효과적이고 빠른 문자 영역 분리 방법을 제안함으로서 처리속도의 향상은 물론 잡영에도 강한 문자 영역 분리 시스템을 구현하였다.