• Current Optics and Photonics
• /
• v.2 no.4
• /
• pp.348-355
• /
• 2018

How to measure the creepage distance of an insulator quickly and accurately is a problem for the power industry at present, and the noticeable concern is that the high voltage insulation equipment cannot be measured online in the charged state. In view of this situation, we develop an on-line measurement system of creepage distance for high voltage insulators based on binocular stereo vision. We have proposed a method of generating linear structured light using a conical off-axis mirror. The feasibility and effect of two ways to solve the interference problem of strong sunlight have been discussed, one way is to use bandpass filters to enhance the contrast ratio of linear structured light in the images, and the other way is to process the images with adaptive threshold segmentation and feature point extraction. After the system is calibrated, we tested the measurement error of the on-line measurement system with a composite insulator sample. Experimental results show that the maximum relative error is 1.45% and the average relative error is 0.69%, which satisfies the task requirement of not more than 5% of the maximum relative error.

• Proceedings of the IEEK Conference
• /
• 2000.06d
• /
• pp.163-166
• /
• 2000

Real-time traffic detection scheme based on Computer Vision is capable of efficient traffic control using automatically computed traffic information and obstacle detection in moving automobiles. Traffic information is extracted by segmenting vehicle region from road images, in traffic detection system. In this paper, we propose the advanced segmentation of vehicle from road images using multiple local region information. Because multiple local region overlapped in the same lane is processed sequentially from small, the traffic detection error can be corrected.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.57 no.7
• /
• pp.1290-1294
• /
• 2008

We present a novel deblurring algorithm for bi-level images blurred by some parameterizable point spread function. The proposed method iteratively searches unknown parameters in the point spread function and noise-to-signal ratio by minimizing an objective function that is based on the binariness and the difference between two intensity values of restoring image. In simulations and experiments, the proposed method showed improved performance compared with the Wiener filtering based method in terms of bit error rate after segmentation.

• The Journal of Fisheries Business Administration
• /
• v.35 no.1
• /
• pp.23-56
• /
• 2004

The offshore trawl fishery is seeking its survival way to overcome current management conditions in red, resulted from the bilateral agreement with China and Japan. However, this movement magnifies conflicts between fisheries on the contrary and it is thought to be impossible to get over current situations. For all that, this study is aimed at investigating how this current situations have occurred. The management freedom as response to the change in fishing conditions of a certain fishery, in case of Korea, is affected by institutional regulations. The example of this is controls on fishing gears, fishing vessels, and fishing grounds. The most exposure of this control is a segmentation of institutional fisheries. The initial segmentation of the offshore trawl fishery in Korea was occurred in the period of Japan's colonization when the degree of use of fishing grounds was limited geographically. At that time, fisheries were divided by fishing areas, but it did not divide the fishery itself. The large - sized fishing vessels were developed politically to be more competative to Japanese fishing vessels since 1950s. During this time, the trawl fishery was merged into current Eastern trawl fishery and South - Western trawl fishery. It was also inevitable to divide into the pair trawl and single trawl fishery as a result of the physical mergency between Western trawl and Southern trawl fishery. In order to develop the trawl fishery, new licenses were issued on the shrimp trawl fishery, through which it was expected to boost the trawl fishery. As opposed, the shrimp trawl fishery was changed into the mid - sized trawl fishery, targeting on the eastern fishing areas and the large - sized trawl fishery was developed since the late of 1970s with the development of filefish processing industry. The large trawl fishery that led in development of offshore trawl fishery since the late of 1950s was started to divide into a pair trawl and single trawl according to the fishing method and capital power. It finally became an institutionally independent fishery in 1980s, respectively. Looking into these historical process, the segmentation of the trawl fishery is thought as a result of the lack of long - term perspective and as a production of trial and error resulted by unprepared policy. As a result, these segmentation of fisheries roles as critical obstacles in harmonization of fisheries and in overcoming of current situations. Therefore, the review of this institutional segmentation of the offshore trawl fishery should be taken for an optimal redistribution of fishing grounds suits with business and fishing technology. For this, the fishery must be divided into large capitalized fishery and small - mid fishery with consideration of capital, fishing method, and the condition of use of fishing grounds. In addition to this, by limiting outline of fishing ground that the large fishery can harvest, it must allow for the small - mid fishery to catch with its own boundary. Furthermore, by launching buyback programs on the trawl, eastern trawl, pair trawl, it can provide broader fishing grounds where the fishery can harvest with management freedom.

• The Journal of the Acoustical Society of Korea
• /
• v.15 no.4E
• /
• pp.21-27
• /
• 1996

In this paper, we propose a discriminative training algorithm for the stochastic segment model (SSM) in continuous speech recognition. As the SSM is usually trained by maximum likelihood estimation (MLE), a discriminative training algorithm is required to improve the recognition performance. Since the SSM does not assume the conditional independence of observation sequence as is done in hidden Markov models (HMMs), the search space for decoding an unknown input utterance is increased considerably. To reduce the computational complexity and starch space amount in an iterative training algorithm for discriminative SSMs, a hybrid architecture of SSMs and HMMs is programming using HMMs. Given the segment boundaries, the parameters of the SSM are discriminatively trained by the minimum error classification criterion based on a generalized probabilistic descent (GPD) method. With the discriminative training of the SSM, the word error rate is reduced by 17% compared with the MLE-trained SSM in speaker-independent continuous speech recognition.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.22 no.4
• /
• pp.771-781
• /
• 1997

In order to solve the problem of uncovered background region due to the region-due to the region-based motion estimation, this paper presents a new method which generates the uncovered region memory using motion estimation and shows the application of the algorithm for very low bit rate video coding. The proposed algorithm can be briefly described as follows it detects the changed region by using the information of FD(frame difference) and segmentation, and then as for only that region the backward motion estimation without transmission of shape information is done. Therefore, from only motion information the uncovered background region memory is generated and updated. The contents stored in the uncovered background region memory are referred whenever the uncovered region comes into existence. The regions with large prediction error are transformed and coded by using DCT. As results of simulation, the proposed algorithm shows the superior improvement in the subjective and objective image quality due to the remarkable reduction of transmission bits for prediction error.

• Journal of Korea Multimedia Society
• /
• v.12 no.3
• /
• pp.383-396
• /
• 2009

Bone age assessment has been widely used in pediatrics to identify endocrine problems of children. Since the number of trained doctors is far less than the demands, there has been numerous requests for automatic estimation of bone age. Therefore, in this paper, we propose an automatic bone age assessment method that utilizes pattern classification techniques. The proposed method consists of three modules; a finger segmentation module, a normalized shape model generation module and a bone age estimation module. The finger segmentation module segments fingers and epiphyseal regions by means of various image processing algorithms. The shape model abstraction module employ ASM to improves the accuracy of feature extraction for bone age estimation. In addition, SVM is used for estimation of bone age. Features for the estimation include the length of bone and the ratios of bone length. We evaluated the performance of the proposed method through statistical analysis by comparing the bone age assessment results by clinical experts and the proposed automatic method. Through the experimental results, the mean error of the assessment was 0.679 year, which was better than the average error acceptable in clinical practice.

• Journal of the Institute of Convergence Signal Processing
• /
• v.3 no.1
• /
• pp.33-40
• /
• 2002

In this paper, we proposed a lossless compression algorithm of medical images which is essential technique in picture archive and communication system. Mammographic image and magnetic resonance image in among medical images used in this study, proposed a region growing segmentation algorithm for compression of these images. A proposed algorithm was partition by three sub region which error image, discontinuity index map, high order bit data from original image. And generated discontinuity index image data and error image which apply to a region growing algorithm are compressed using JBIG(Joint Bi-level Image experts Group) algorithm that is international hi-level image compression standard and proper image compression technique of gray code digital Images. The proposed lossless compression method resulted in, on the average, lossless compression to about 73.14% with a database of high-resolution digital mammography images. In comparison with direct coding by JBIG, JPEG, and Lempel-Ziv coding methods, the proposed method performed better by 3.7%, 7.9% and 23.6% on the database used.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.19 no.4
• /
• pp.701-708
• /
• 2024

The advent of transformer architectures in deep learning has been a major breakthrough in natural language processing research. Object name recognition is a branch of natural language processing and is an important research area for tasks such as information retrieval. It is also important in the biomedical field, but the lack of Korean biomedical corpora for training has limited the development of Korean clinical research using AI. In this study, we built a new biomedical corpus for Korean biomedical entity name recognition and selected language models pre-trained on a large Korean corpus for transfer learning. We compared the name recognition performance of the selected language models by F1-score and the recognition rate by tag, and analyzed the errors. In terms of recognition performance, KlueRoBERTa showed relatively good performance. The error analysis of the tagging process shows that the recognition performance of Disease is excellent, but Body and Treatment are relatively low. This is due to over-segmentation and under-segmentation that fails to properly categorize entity names based on context, and it will be necessary to build a more precise morphological analyzer and a rich lexicon to compensate for the incorrect tagging.

• ETRI Journal
• /
• v.33 no.3
• /
• pp.423-433
• /
• 2011

This paper introduces an adaptive and integrated utterance verification (UV) framework using minimum verification error (MVE) training as a new set of solutions suitable for real applications. UV is traditionally considered an add-on procedure to automatic speech recognition (ASR) and thus treated separately from the ASR system model design. This traditional two-stage approach often fails to cope with a wide range of variations, such as a new speaker or a new environment which is not matched with the original speaker population or the original acoustic environment that the ASR system is trained on. In this paper, we propose an integrated solution to enhance the overall UV system performance in such real applications. The integration is accomplished by adapting and merging the target model for UV with the acoustic model for ASR based on the common MVE principle at each iteration in the recognition stage. The proposed iterative procedure for UV model adaptation also involves revision of the data segmentation and the decoded hypotheses. Under this new framework, remarkable enhancement in not only recognition performance, but also verification performance has been obtained.