• 정보관리학회지
• /
• 제35권2호
• /
• pp.37-62
• /
• 2018

문헌정보학 분야의 국내 학술지 논문으로 구성된 문헌집합을 대상으로 기계학습에 기초한 자동분류의 성능에 영향을 미치는 요소들을 검토하였다. 특히, "정보관리학회지"에 수록된 논문에 주제 범주를 자동 할당하는 분류 성능 측면에서 용어 가중치부여 기법, 학습집합 크기, 분류 알고리즘, 범주 할당 방법 등 주요 요소들의 특성을 다각적인 실험을 통해 살펴보았다. 결과적으로 분류 환경 및 문헌집합의 특성에 따라 각 요소를 적절하게 적용하는 것이 효과적이며, 보다 단순한 모델의 사용으로 상당히 좋은 수준의 성능을 도출할 수 있었다. 또한, 국내 학술지 논문의 분류는 특정 논문에 하나 이상의 범주를 할당하는 복수-범주 분류(multi-label classification)가 실제 환경에 부합한다고 할 수 있다. 따라서 이러한 환경을 고려하여 단순하고 빠른 분류 알고리즘과 소규모의 학습집합을 사용하는 최적의 분류 모델을 제안하였다.

• 정보관리학회지
• /
• 제33권2호
• /
• pp.33-59
• /
• 2016

국내 학술회의 논문으로 구성된 문헌집합을 대상으로 기계학습에 기초한 자동분류의 성능에 영향을 미치는 요소들을 검토하였다. 특히 구현이 쉽고 컴퓨터 처리 속도가 빠른 로치오 알고리즘을 사용하여 "한국정보관리학회 학술대회 논문집"의 논문에 주제 범주를 자동 할당하는 분류 성능 측면에서 분류기 생성 방법, 학습집합 규모, 가중치부여 기법, 범주 할당 방법 등 주요 요소들의 특성을 다각적인 실험을 통해 살펴보았다. 결과적으로 분류 환경 및 문헌집합의 특성에 따라 파라미터(${\beta}$, ${\lambda}$)와 학습집합의 크기(5년 이상)를 적절하게 적용하는 것이 효과적이며, 동등한 성능 수준이라면 보다 단순한 단일 가중치부여 기법을 사용하여 분류의 효율성을 높일 수 있음을 발견하였다. 또한 국내 학술회의 논문의 분류는 특정 논문에 하나 이상의 범주가 부여되는 복수-범주 분류(multi-label classification)가 실제 환경에 부합한다고 할 수 있으므로, 이러한 환경을 고려하여 주요 성능 요소들의 특성에 기초한 최적의 분류 모델을 개발할 필요가 있다.

• 제어로봇시스템학회논문지
• /
• 제22권10호
• /
• pp.839-846
• /
• 2016

Bin picking is a task of picking a small object from a bin. For accurate bin picking, the 3D pose information, position, and orientation of a small object is required because the object is mixed with other objects of the same type in the bin. Using this 3D pose information, a robotic gripper can pick an object using exact distance and orientation measurements. In this paper, we propose a 3D vision technique for accurate measurement of 3D position and orientation of small objects, on which a paper label is stuck to the surface. We use a maximally stable extremal regions (MSERs) algorithm to detect the label areas in a left bin image acquired from a stereo camera. In each label area, image features are detected and their correlation with a right image is determined by a stereo vision technique. Then, the 3D position and orientation of the objects are measured accurately using a transformation from the camera coordinate system to the new label coordinate system. For stable measurement during a bin picking task, the pose information is filtered by averaging at fixed time intervals. Our experimental results indicate that the proposed technique yields pose accuracy between 0.4~0.5mm in positional measurements and $0.2-0.6^{\circ}$ in angle measurements.

• Journal of Electrical Engineering and Technology
• /
• 제6권4호
• /
• pp.551-555
• /
• 2011

The Crab label tag with a loop matching feed and a modified dipole antenna structure was proposed. The antenna impedance is conjugated easily to a radio frequency identification IC chip impedance by a loop matching feed. The reading range of the crab structure tag is 0.9-1.0 m from the upper side of the formula milk can lid. The fabricated label tag size is $44.0{\times}44.0mm^2$. The operating frequency at -3 dB return loss is 861.0-929.0 MHz, and the maximum reading range at the anechoic chamber is 1.5 m.

• 한국멀티미디어학회논문지
• /
• 제23권6호
• /
• pp.756-764
• /
• 2020

A typical algorithm used for image analysis is the Convolutional Neural Network(CNN). R-CNN, Fast R-CNN, Faster R-CNN, etc. have been studied to improve the performance of the CNN, but they essentially require large amounts of data and high algorithmic complexity., making them inappropriate for small and medium-sized services. Therefore, in this paper, the image label prediction algorithm based on CNN with collaborative layer with low complexity, high accuracy, and small amount of data was proposed. The proposed algorithm was designed to replace the part of the neural network that is performed to predict the final label in the existing deep learning algorithm by implementing collaborative filtering as a layer. It is expected that the proposed algorithm can contribute greatly to small and medium-sized content services that is unsuitable to apply the existing deep learning algorithm with high complexity and high server cost.

• 정보처리학회논문지A
• /
• 제8A권2호
• /
• pp.179-188
• /
• 2001

We analyze the property of candidate node set in the network graph, and propose an algorithm to decrease shortest path-finding computation time by using multiple dynamic-range queue(MDRQ) structure. This MDRQ structure is newly created for effective management of the candidate node set. The MDRQ algorithm is the shortest path-finding algorithm that varies range and size of queue to be used in managing candidate node set, in considering the properties that distribution of candidate node set is constant and size of candidate node set rapidly change. This algorithm belongs to label-correcting algorithm class. Nevertheless, because re-entering of candidate node can be decreased, the shortest path-finding computation time is noticeably decreased. Through the experiment, the MDRQ algorithm is same or superior to the other label-correcting algorithms in the graph which re-entering of candidate node didn’t frequently happened. Moreover the MDRQ algorithm is superior to the other label-correcting algorithms and is about 20 percent superior to the other label-setting algorithms in the graph which re-entering of candidate node frequently happened.

• 정보교육학회논문지
• /
• 제11권2호
• /
• pp.185-194
• /
• 2007

유비쿼터스 사회의 도래에 따라 교육의 패러다임도 변화되고 있다. 유비쿼터스 환경의 u-러닝에서는 휴대폰, PDA, PMP, UMPC, TPC 등의 모바일 기기가 중요한 학습도구가 될 것이다. 특히, 유비쿼터스 사회의 핵심 기술인 RFID를 모바일 기기와 함께 교수-학습에 활용한다면 의미있는 학습이 되기에 충분하다. 본 연구에서는 RFID를 교수-학습에 적용하기 위한 방안으로 RFID-card, RFID-book, RFID-label, RFID-test의 네 가지를 제시하였으며, RFID-label의 방법을 실제 교수-학습에 적용하고 그 가능성과 효과성을 탐색하였다. 그 결과 RFID의 교수-학습 적용에 대한 긍정적인 효과와 가능성을 얻었다. 하지만 RFID를 교수-학습에 적용함에 있어서 기술적으로 부족한 점이 나타났으며, RFID의 교수-학습 적용을 위한 모델의 지속적인 연구의 필요성이 제기되었다.

• 대한기계학회논문집A
• /
• 제25권5호
• /
• pp.841-849
• /
• 2001

The pixel information of the object was obtained sequentially and pixels were clustered to a label by the line labeling method. Feature points were determined by finding the slope for edge pixels after selecting the fixed number of edge pixels. The slope was estimated by the least square method to reduce the detection error. Once a matching point was determined by comparing the feature information of the object and the pattern, the parameters for translation, scaling and rotation were obtained by selecting the longer line of the two which passed through the matching point from left and right sides. Finally, modified Hausdorff Distance has been used to identify the similarity between the object and the given pattern. The multi-label method was developed for recognizing the patterns with more than one label, which performs the modified Hausdorff Distance twice. Experiments have been performed to verify the performance of the proposed algorithm and method for simple target image, complex target image, simple pattern, and complex pattern as well as the partially hidden object. It was proved via experiments that the proposed image matching algorithm for recognizing the object had a good performance of matching.

• 한국비블리아학회지
• /
• 제4권1호
• /
• pp.85-102
• /
• 1980

Accepting the necessity for maintaining the objectives of the existing circulation system, the computer-based system could be designed by the system analyst and librarians to gain a variety of improvements in the maintenance, accessibility of circulation records and more meaningful statistical records. If the terminal can be operated on-line, then this circulation data is transmitted directly to the computer, where it may update to the circulation file immediately or alternatively be kept in direct access file for updating in batch mode. on-line system in the circulation operations is "data-collection system" and "Bar-coded label system" Bar-coded label system is simple, quick, and error-free input of data. Attached to CRT terminal is a "light pen" which is hand held and will read a bar-coded label as the pen is passed over the labels (one affixed to the book itself, other carried on the borrower cards). Instantaneously the data concerning transaction is stored in the central mini-computer. It is useful, economical for us to co-operate many libraries in Korea and design borrower's ID code, book no., classification code in the Bar-coded label system by the members of the computer center and the library staff at every stage. As for book loan, the borrowers ID code, book number and classification code are scanned by the bar-code scanner or light pen and the computer decides whether to loan and store the data. The visual display unit shows the present status of a borrowers borrowing and decides whether borrower can borrow.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2018년도 추계학술대회
• /
• pp.201-202
• /
• 2018

본 논문에서는 기존 ImageNet과 Oxford-IIIT Pet Image의 Dataset에서 제공하는 개 품종 이미지와 인터넷 상에서 개 품종 이미지를 데이터 마이닝을 통해 획득된 개 품종 이미지를 결합하고 Random-Label을 추가 하여 개 품종 122개의 클래스와 개 품종이 아닌 1개의 클래스를 인식하는 방법에 대해 소개 한다. 기존 DB만을 사용하였을 때 개 품종 인식률 대비 기존 DB와 수집 DB를 모두 사용한 개 품종 인식률이 Top-1에 대해서 1.5% 개선되었다. 개가 아닌 이미지 인식은 랜덤 DB를 10000장의 경우 93% 인식률을 확인했다.