• Journal of Multimedia Information System
• /
• v.6 no.4
• /
• pp.225-234
• /
• 2019

Deep learning has become increasingly popular in both academic and industrial areas nowadays. Various domains including pattern recognition, Computer vision have witnessed the great power of deep neural networks. However, current studies on deep learning mainly focus on quality data sets with balanced class labels, while training on bad and imbalanced data set have been providing great challenges for classification tasks. We propose in this paper a method of data analysis-based data reduction techniques for selecting good and diversity data samples from a large dataset for a deep learning model. Furthermore, data sampling techniques could be applied to decrease the large size of raw data by retrieving its useful knowledge as representatives. Therefore, instead of dealing with large size of raw data, we can use some data reduction techniques to sample data without losing important information. We group PCB characters in classes and train deep learning on the ResNet56 v2 and SENet model in order to improve the classification performance of optical character recognition (OCR) character classifier.

• Journal of Korea Society of Industrial Information Systems
• /
• v.13 no.1
• /
• pp.22-29
• /
• 2008

This paper describes the developed results of vision system for the recognition of material management characters, which was included in the billet image. The material management characters, which was marked at the surface of billet, should be recognized before billet moves to the next process. Our vision system for the character recognition includes that CCD camera system which acquire billet image, optical transmission system which transmit captured image to the long distance, input and output system for the interface with existing system and software for the character recognition. We have installed our vision system at the wire rod line of steel & iron plant and tested. Also, we have performed inspection of durability, reliability and recognition rate. Through the testing, we have confirmed that our system have high recognition rate, 98.6%.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.14 no.1
• /
• pp.20-27
• /
• 2004

We propose the character recognition system for Gerber files. The Gerber file is the vector-formatted drawing file for PCB manufacturing, which includes various symbols, figures and characters. Also, the characters are written in horizontal, vertical, and reverse-vortical directions. In this paper, we newly propose the Gerber-character recognition system to recognize all of component names located in PCB. To improve the performance, we develop the multiple recognizers by neural networks and the verifier considering the structural features. The developed system has been installed to the auto-programming software for PCB assembly and inspection machines.

• Journal of the Institute of Convergence Signal Processing
• /
• v.5 no.2
• /
• pp.165-171
• /
• 2004

Owing to small sizes and high density to the semiconductor It, it is difficult to discriminate the defects of ICs by human eyes. High precision inspection system with computer vision is essentially established for the manufacturing process due to the variety of defective parts. Especially it is difficult to implement the algorithm for the coplanarity of IC leads. Therefore in this paper, the inspection system which can detect the defects of the SOP types of ICs having 1cm${\times}$0.5cm of the chip size is implemented and evaluated it's performance. In order to optimally detect various items, some principles of geometry are theoretically presented , length measurement, pitch measurement, angle measurement, brightness of image and correcton of position. The interface circuit is designed for implementation of inspection system and connected the HANDLER. In the result, the system could detect two ICs' defects per second and confirmed the resolution of 20$\mu$m per pixel.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.4 no.1
• /
• pp.55-63
• /
• 2008

I Previous scrolling-shooter games cannot represent various emotions of characters as a personality and situations, so it can only represent simple and momentary emotions. This paper purpose to develop and applying an artificial emotion which gives emotional responses and actions to character as a personality and situations of a character. So we proposed an artificial emotion that receives emotional stimulus, analyzes what emotion will be generated, controls various emotions by a personality and time, and exports a current emotion. To visualize and to test an artificial emotion, We made a scrolling shooter game which has two characters who have a different personality each other. And we applied an artificial emotion to that game, modified it to be able to change status as emotions. To estimate a satisfaction of a artificial emotion for a scrolling-shooter game, we made up a question to two groups-one has people who likes a scrolling shooter game, the other has people who doesn't.

• Journal of Korea Multimedia Society
• /
• v.24 no.3
• /
• pp.382-395
• /
• 2021

In this study, character recognition using deep learning is performed among the various defects in the PCB, the purpose of which is to check whether the printed characters are printed correctly on top of components, or the incorrect parts are attached. Generally, character recognition may be perceived as not a difficult problem when considering MNIST, but the printed letters on the PCB component data are difficult to collect, and have very high redundancy. So if a deep learning model is trained with original data without any preprocessing, it can lead to over fitting problems. Therefore, this study aims to reduce the redundancy to the smallest dataset that can represent large amounts of data collected in limited production sites, and to create datasets through data enhancement to train a flexible deep learning model can be used in various production sites. Moreover, ResNet model verifies to determine which combination of datasets is the most effective. This study discusses how to reduce and augment data that is constantly occurring in real PCB production lines, and discusses how to select coresets to learn and apply deep learning models in real sites.

• Journal of Conservation Science
• /
• v.27 no.1
• /
• pp.123-126
• /
• 2011

The results of the inspection of the vase condition for April 29th, 2008 have stated the mineral-petrographic characteristic of the material, the character of the stone mechanical defects; its impurity and biological faults have been described; the results of ultrasonic sounding of the stone have been shown. Possible reasons and process of destruction of the stone vase have been analysed on January, 6 to 12th, 2008. Recommendations on restoration and preservation of the vase have been given.

• Journal of Korean Institute of Industrial Engineers
• /
• v.5 no.1
• /
• pp.19-25
• /
• 1979

This paper presents an analysis of the contents of job for the sensory inspection on the basis of the probability theory, and the new determination for an index(d') of the testing power precision in carrying out sensory inspections. Also presented are the evaluation method of determining the ability of inspector by presuming the confidence interval for the average record of inspector, and the computation method for the index (de') of the testing power precision required as the goal-value in accordance with quality character, process inferior ratio, and required AOQ.

• Laser Solutions
• /
• v.17 no.3
• /
• pp.10-14
• /
• 2014

By the specific character of femtosecond laser controlled volume of magnitude of micrometer scale could be ablated without melting phase in SKD11 and SUS304. According to the laser machining parameters various sectional shapes could be engraved on the surface of metals. Typical engraved lines were $10{\mu}m$ wide and deep. Coarse-milled surface was made $10{\mu}m$ lower than the original elevation by a bunch of laser-engraved lines in suitable spacing. The repeated banks with a height of $10{\mu}m$ could be made with the combination of the intact area.

• Journal of the Korea institute for structural maintenance and inspection
• /
• v.10 no.1
• /
• pp.216-222
• /
• 2006

This study is a simulative Paper on scour around piers. As materials on the river bed have been changing through erosion, transportation and accumulation, so does the waterway's section. Analyzing them through the BOSS SMS. this study deals with the local-scour characters with the change of shapes in row piers which have been built in the identical direction of water flow.