• Smart Structures and Systems
• /
• v.29 no.1
• /
• pp.207-220
• /
• 2022

Structural health monitoring (SHM) plays a vital role in the maintenance and operation of constructions. In recent years, autonomous inspection has received considerable attention because conventional monitoring methods are inefficient and expensive to some extent. To develop autonomous inspection, a potential approach of crack identification is needed to locate defects. Therefore, this study exploits two deep learning-based segmentation models, DeepLabv3+ and Mask R-CNN, for crack segmentation because these two segmentation models can outperform other similar models on public datasets. Additionally, impacts of label quality on model performance are explored to obtain an empirical guideline on the preparation of image datasets. The influence of image cropping and label refining are also investigated, and different strategies are applied to the dataset, resulting in six alternated datasets. By conducting experiments with these datasets, the highest mean Intersection-over-Union (mIoU), 75%, is achieved by Mask R-CNN. The rise in the percentage of annotations by image cropping improves model performance while the label refining has opposite effects on the two models. As the label refining results in fewer error annotations of cracks, this modification enhances the performance of DeepLabv3+. Instead, the performance of Mask R-CNN decreases because fragmented annotations may mistake an instance as multiple instances. To sum up, both DeepLabv3+ and Mask R-CNN are capable of crack identification, and an empirical guideline on the data preparation is presented to strengthen identification successfulness via image cropping and label refining.

• 제어로봇시스템학회:학술대회논문집
• /
• 1989.10a
• /
• pp.1005-1010
• /
• 1989

A refined version of automatic micro-organism recognition and identification method, 'O.I.S.M.2' is proposed in this paper, using image processing based on an expert system. This proposed method is based on the segmentation of the organism image, characterizing segment features, which are independent of individual size and length. Complicated shapes of organisms are divided into basic shape segments defined in this paper such as lines, circles, ovals etc. Organisms can then be expressed simply in a set of segments, regardless their individual differences.

• Proceedings of the IEEK Conference
• /
• 2001.06b
• /
• pp.75-78
• /
• 2001

In this paper, we proposed a personal identification method using binay image encryption technique and decryption system in JTC structure. Logo which represents the group symbol was encrypted with personal fingerprint and JTC structure decrypts this logo. The logo can not decrypted by other unused fingerprint even if the encrypted image was lost or stolen. So this method can give more safe personal identification.

• International journal of advanced smart convergence
• /
• v.4 no.2
• /
• pp.84-93
• /
• 2015

This study is for the development of Cover Glass Grinding Processing System. This system is developed for manufacturing a mass product system grinding cover glasses with highly precise mechanism, and we improved resulted quality. In the development process, we developed a complete process technology through mechanical design, image processing technology, spindle control, mark identification algorithm etc. With this cover glass grinding development, we could developed process technology, image processing technology, organization mechanisms and control algorithms.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.12
• /
• pp.5087-5102
• /
• 2015

Rapid popularization of smart cell phone equipped with camera has led to a number of new legal and criminal problems related to multimedia such as digital image, which makes cell phone source identification an important branch of digital image forensics. This paper proposes a classifier combination based source identification strategy for cell phone images. To identify the outlier cell phone models of the training sets in multi-class classifier, a one-class classifier is orderly used in the framework. Feature vectors including color filter array (CFA) interpolation coefficients estimation and multi-feature fusion is employed to verify the effectiveness of the classifier combination strategy. Experimental results demonstrate that for different feature sets, our method presents high accuracy of source identification both for the cell phone in the training sets and the outliers.

• Journal of the Korea Convergence Society
• /
• v.9 no.2
• /
• pp.295-302
• /
• 2018

The purpose of this study was to empirically examine how the perception of CSR of domestic professional baseball clubs was related to the reputation of clubs, identification of clubs, and parent company image. For the research, a survey was conducted by 277 fans of LG Twins, Doosan Bears, SK Wyverns, and KT Wiz in the metropolitan area. First, the reputation of the club had a positive impact on club reputation. Second, CSR of professional baseball clubs had a positive impact on club identification. Third, the reputation of the club had a positive impact on the image of the professional baseball team. Fourth, the identification of clubs had a positive effect on the image of the professional baseball team's mother-company.

• Journal of Digital Convergence
• /
• v.17 no.9
• /
• pp.105-114
• /
• 2019

This study examined the causal relationship between CSV and brand image and consumers' behavioral intentions by structural equation model using psychological variables such as CSV intentionality and company-consumer identification. 198 questionnaires were used to test 8 hypotheses using Smart PLS 2.0. Results are as follows: First, two hypotheses were rejected that CSV had a negative effect on CSV intentionality and CSV intentionality had a negative effect on brand image. Second, CSV has a positive effect on company-consumer identification and company-consumer identification has a positive effect on brand image, purchase intention, and willingness to pay. This study suggests the necessity of enhancing customer participation and communication for the company to establish CSV strategies.

• Journal of the Semiconductor & Display Technology
• /
• v.18 no.3
• /
• pp.31-35
• /
• 2019

Human beings have been fascinated by the applicability of the medium of photography since the device was first introduced in the thirteenth century to acquire images by attempting primitive and rudimentary approaches. In the 21st century, it has been developed as a wide range of technology that enables not only the application of artistic expression as a method of replacing the human-hand-painted screen but also the planar recording form in the format of video or image. It is more effective to use the information extracted from the image data rather than to use a randomly given file name in order to provide a variety of services in the offline or online system. When extracting an identifier from a region of an image, high cost cannot be avoided. This paper discusses the image entropy-based approach and proposes a linear methodology to measure the image entropy in an effort to devise a solution to this method.

• Journal of the Optical Society of Korea
• /
• v.4 no.1
• /
• pp.19-22
• /
• 2000

A new image encoding and identification scheme is proposed for security verification by us-ing a CGH(computer generated hologram), random phase mask, and a correlation technique. The encrypted image, which is attached to the security product, is made by multiplying a QP- CGH(quadratic phase CGI) with a random phase function. The random phase function plays a key role when the encrypted image is decrypted. The encrypted image can be optically recovered by a 2-f imaging system and automatically verified for personal identification by a 4-f correlation system. Simulation results show the proposed method can be used for both the reconstruction of an original image and the recognition of an encrypted image.

• Journal of the Semiconductor & Display Technology
• /
• v.16 no.1
• /
• pp.6-10
• /
• 2017

Recent enhancement in the still image acquisition devices has been widely perpetrated into the daily life of the common people. Due to this trend, the voluminous still images, that are produced and shared in the personal or the massive storage, need to controlled with effective and efficient management. The human-devised or system-generated still image identifiers used for the identification of the images are at risk in the situation of unexpected changing or eliminating of the identifiers. In this paper, we propose a parallel processing based method for still image identifier generation by utilizing the still image internal features.