• Journal of KIISE:Computing Practices and Letters
• /
• v.15 no.5
• /
• pp.350-354
• /
• 2009

To optimize the performance of software programs, it is important to know certain hardware parameters such as the CPU speed, the cache size, the number of TLB entries, and the parameters of the memory subsystem. There exist several ways to obtain the values of various hardware parameters. Firstly. the values can be taken from the hardware manual. Secondly, the parameters can be obtained by calling functions provided by the operating systems. Finally, hardware detection programs can find the desired values. Such programs are usually executed on PC or server systems and report the CPU speed, the cache size, the number of TLB entries, and so on. However, they do not sufficiently detect the parameters of one of the most important parts of the computer concerning performance, namely the memory bank layout in the memory subsystem. In this paper, we present an algorithm to detect the memory bank parameters. We run an implementation of our algorithm on various embedded systems and compare the detected values with the real hardware parameters. The results show that the presented algorithm detects the cache size, the number of TLB entries, and the memory bank layout with high accuracy.

• Journal of Korea Multimedia Society
• /
• v.16 no.4
• /
• pp.430-443
• /
• 2013

The recognition of current position and road surrounding of car is very important to driver for safe driving. This paper presents the recognition technique of the road traveling environment using position information hiding and viewpoint transform that monitors the information of road furniture and signature and notifies them to driver. The proposed scheme generates the road images into which the position information are hided, from car camera and GPS module and provides the road information to driver through the viewpoint transformation and the road signature detection. The driving tests with camera and GPS module verified that the position information hiding takes about 66.5ms per frame, the detection rate of road signature is about 95.83%, and the road signature detection takes about 227.45ms per frame. Therefore, we know that the proposed scheme can recognize the road traveling environment on the road video with 15 frame rate.

• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.1
• /
• pp.27-36
• /
• 2019

In this paper, we propose a method to find the most similar answer to the user's response from the question-answer database in order to avoid generating a redundant question in retrieval-based automatic question generation system. As a question of the most similar answer to user's response may already be known to the user, the question should be removed from a set of question candidates. A similarity detector calculates a similarity between two answers by utilizing the same words, paraphrases, and sentential meanings. Paraphrases can be acquired by building a phrase table used in a statistical machine translation. A sentential meaning's similarity of two answers is calculated by an attention-based convolutional neural network. We evaluate the accuracy of the similarity detector on an evaluation set with 100 answers, and can get the 71% Mean Reciprocal Rank (MRR) score.

• Journal of the Microelectronics and Packaging Society
• /
• v.8 no.3
• /
• pp.1-9
• /
• 2001

This paper proposes an inspection algorithm for micro ball grid array ($\mu$BGA) solder balls. This algorithm is motivated by the difficulty of finding defect balls by human visual inspection due to their small dimensions. Specifically, it is developed herein an automated vision-based inspection algorithm for $\mu$BGA's, which can inspect solder balls not only for so-called two dimensional errors, such as missings, positions and sizes, but also for height errors. The inspection algorithm uses two dimensional images of $\mu$BGA obtained through special blue illumination, and processes them with a rotation-invariant sub algorithm. It can also detect height errors when a two-camera system is available. Simulation results show that the proposed algorithm is more efficient in detecting ball defects compared with the conventional algorithms.

• Annual Conference of KIPS
• /
• 2023.05a
• /
• pp.448-450
• /
• 2023

본 논문에서는 복부 CT 의료영상에서 근감소증 진단을 위한 지표로 활용하는 요추 3번 슬라이스를 분류하기 위해서 CNN 기반의 EfficientNetV2를 사용하여 자동분류모델을 개발하였다. 이를 위해 먼저 전체 복부 CT 의료영상에서 Thoracic, L1, L2, L3, L4, L5, Sacral 7개의 슬라이스를 검출하도록 하였다. 자동분류모델의 정확성을 측정하기 위해서 Test 데이터셋을 사용하여 Confusion Matrix 결과를 통해 개발된 모델의 성능을 검증한 결과를 보였다. 본 연구결과는 복부 CT 영상에서 기존 L3 레벨의 특정 단면에서 근육량을 측정하는 것에서 다양한 부위에서 측정할 수 있는 장점을 갖게 된다. 그리고 의료영상기반의 근감소증 진단 연구에 도움을 줄 것으로 기대하고 있다.

• Journal of KIISE:Software and Applications
• /
• v.28 no.5
• /
• pp.389-401
• /
• 2001

This paper proposes an abstraction mechanism of the low-level digital video features for the automatic retrievals of the explosion scenes from the digital video library. In the proposed abstraction mechanism, the regional dominant colors of the key frame and the motion energy of the shot are defined as the primary abstractions of the shot for the explosion scene retrievals. It is because an explosion shot usually consists of the frames with a yellow-tone pixel and the objects in the shot are moved rapidly. The regional dominant colors of shot are selected by dividing its key frame image into several regions and extracting their regional dominant colors, and the motion energy of the shot is defined as the edge image differences between key frame and its neighboring frame. The edge image of the key frame makes the retrieval of the explosion scene more precisely, because the flames usually veils all other objects in the shot so that the edge image of the key frame comes to be simple enough in the explosion shot. The proposed automatic retrieval algorithm declares an explosion scene if it has a shot with a yellow regional dominant color and its motion energy is several times higher than the average motion energy of the shots in that scene. The edge image of the key frame is also used to filter out the false detection. Upon the extensive exporimental results, we could argue that the recall and precision of the proposed abstraction and detecting algorithm are about 0.8, and also found that they are not sensitive to the thresholds. This abstraction mechanism could be used to summarize the long action videos, and extract a high level semantic information from digital video archive.

• Journal of Korean Ophthalmic Optics Society
• /
• v.21 no.2
• /
• pp.119-126
• /
• 2016

Purpose: To analyze the effect of accommodative control and change values between subjective refraction (SR) and auto-refraction (AR) according to application of fogging after accommodative stimulation depending on ametropia type. Methods: Myopic ametropia 76 eyes and hyperopic ametropia 52 eyes participated for this study. SR and AR values measured by three test conditions (Before accommodative stimulation; Before AS, After accommodative stimulation; After AS, and After application of fogging; After AF) were compared, respectively. Results: In myopic eyes, (-)spherical power by SR and AR in After AS test was significantly increased as compared to Before AS test, (-)spherical power in After AF test was decreased to the level of Before AS test. The differences of spherical power between SR and AR were highly measured by SR in After AS test, and highly measured by AR in After AF test, respectively. In hyperopic eyes, (+)spherical power of SR significantly decreased in After AS test compared to Before AS test, more (+)spherical power was detected in After AF test compared to Before AS test. (+)spherical power of AR have no significant difference between Before AS and After AS test, but more (+)spherical power was detected in After AF test compared to Before AS test. The differences of (+)spherical power between SR and AR were significant in all test conditions. Among 52 eyes which were measured as hyperopic ametropia, 7 eyes were measured as myopia by SR in After AS test. In case of AR, 25 eyes among 52 eyes were mismeasured as myopia of ranges from -0.25 D to -1.25 D in Before AS test, 26 eyes in After AS test, and 19 eyes in After AF test were mismeasured as myopia of ranges from -0.25 D to -1.25 D. Conclusions: Regardless of ametropia type, accommodative control by After AF test was effective on both refraction process. However, in auto-refraction for hyperopic eyes, the misdetermined proportion of refractive error's type was high due to consistent accommodative intervention in all test condition. Therefore, in order to obtain an accurate value of refractive errors, full correction should be determined by subjective refraction process after fogging method.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.37 no.8A
• /
• pp.639-647
• /
• 2012

Vision-based on-road vehicle detection is one of the key techniques in automotive driver assistance systems. However, due to the huge within-class variability in vehicle appearance and environmental changes, it remains a challenging task to develop an accurate and reliable detection system. In general, a vehicle detection system consists of two steps. The candidate locations of vehicles are found in the Hypothesis Generation (HG) step, and the detected locations in the HG step are verified in the Hypothesis Verification (HV) step. Since the final decision is made in the HV step, the HV step is crucial for accurate detection. In this paper, we propose using a reduced multivariate polynomial pattern classifier (RM) for the HV step. Our experimental results show that the RM classifier outperforms the well-known Support Vector Machine (SVM) classifier, particularly in terms of the fast decision speed, which is suitable for real-time implementation.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2016.05a
• /
• pp.823-826
• /
• 2016

This paper proposes a method for detecting automatically a smartphone camera in order to put a security stickers and detecting security sickers to recognize whether the removable security sticker or not. The method reduce the time that a person put a security stickers and check the stickers whether the removable stickers at rush hour. It is significantly low to detect rate if the image quality is poor low by LED light flickering phenomenon at the time of the camera detection of the smartphone. Therefor, in this paper, it implement the method that detecting camera to inspect frequently degree after applying the adaptive threshold. Further, in order to recognize whether the removable security sticker or not implement the sticker region detection method using a color filter. The proposed method could increase detecting ratio than earlier methods with LED light flickering phenomenon. In addition, it was also high security sticker detection ratio.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.11 no.1
• /
• pp.107-112
• /
• 2016

Fastener is a very important component that is used in various areas in industry. Recently, various color fasteners are introduced. According to this, online inspection is required in this area. In this paper, an algorithm for the automatic extraction of the surface of color fastener using color information and dynamic programming is presented. The outer boundary of fastener is found using the difference of color that enables robust processing. The inner boundary of fastener is found by dynamic programming that uses the difference of brightness value within fixed area after converting image to polar coordinate. Experiments are done using the same parameters.