• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2005.05a
• /
• pp.77-86
• /
• 2005

컨테이너 터미널에서는 장치장으로 반입되는 수출 컨테이너의 무게를 몇 단계 그룹으로 나누고 각 무게그룹 별로 모아서 장치한다. 이는 수출 컨테이너를 선박에 싣는 적하 작업 시 선박의 안정성을 위하여 무거운 무게그룹의 컨테이너들을 장치장에서 먼저 반출하여 선박의 바닥 쪽에 놓기 위함이다. 하지만 반입되는 컨테이너의 무게그룹을 결정할 때 사용하는 운송사로부터 받은 무게정보는 부정확한 경우가 많아 하나의 스택(stack)에 서로 다른 무게그룹에 속하는 컨테이너들이 섞이게 된다. 이로 인하여 무거운 무게그룹의 컨테이너를 반출할 때 해당 컨테이너의 상단에 놓여진 보다 가벼운 무게그룹의 컨테이너들을 임시로 옮겨야 하는 재취급(rehandling, reshuffling)이 발생하게 된다. 적하작업 시 장치장에서 재취급이 빈번히 발생하면 작업이 지연되므로 터미널 생산성 향상을 위해서는 재취급 발생을 가급적 줄여야 한다. 본 논문에서는 기계학습 기법을 적용하여 반입 컨테이너의 무게그룹을 보다 정확히 추정하는 방안을 제안한다. 또한 탐색을 통하여 분류기 생성에 관여하는 비용행렬(cost matrix)을 조정함으로써 재취급 발생을 줄일 수 있는 분류기(classifier)를 생성하는 방안을 함께 소개한다. 실험 결과 본 논문에서 제안하는 방안 적용 시 재취급 발생을 $5{\sim}7%$ 정도 줄일 수 있음을 예상할 수 있었다.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2018.07a
• /
• pp.339-340
• /
• 2018

산업 현장에서 사용되는 산업용 로봇은 자동화 시스템 도입에 따라 큰 성장세를 보이고 있다. 본 논문에서는 컨베이어 벨트에 올라온 물체의 무게를 측정하고, 로봇 팔이 무게별로 물건을 지정 위치에 분류하는 시스템을 구현하였다. 또 로봇 팔끼리 서로 데이터를 주고받으며 데이터 값에 따른 동작을 수행하고 클라우드에 데이터 값을 업로드하여 쉽게 모니터링 가능한 로봇 팔을 이용한 무게 분류 시스템을 설계하였다.

• Journal of Intelligence and Information Systems
• /
• v.11 no.2
• /
• pp.59-79
• /
• 2005

Export containers in a container terminal are usually classified into a few weight groups and those belonging to the same group are placed together on a same stack. The reason for this stacking by weight groups is that it becomes easy to have the heavier containers be loaded onto a ship before the lighter ones, which is important for the balancing of the ship. However, since the weight information available at the time of container arrival is only an estimate, those belonging to different weight groups are often stored together on a same stack. This becomes the cause of extra moves, or rehandlings, of containers at the time of loading to fetch out the heavier containers placed under the lighter ones. In this paper, we use machine learning techniques to derive a classifier that can classify the containers into the weight groups with improved accuracy. We also show that a more useful classifier can be derived by applying a cost-sensitive learning technique, for which we introduce a scheme of searching for a good cost matrix. Simulation experiments have shown that our proposed method can reduce about 5$\sim$7% of rehandlings when compared to the traditional weight grouping method.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.10a
• /
• pp.229-232
• /
• 2022

Recently, production of offshore fisheries in Korea has been decreasing. Since production of offshore fisheries in 2016 fell below 1 million tons for the first time in 44 years, it has not recovered and has been decreasing. In order to cope with such a decrease in fishery resources, the TAC (total allowable catch) system is implemented internationally for fisheries resource management. Since 1999, South Korea has introduced the TAC system to perform resource management. In this paper, we propose an artificial intelligence-based fish classification and weight estimation system that can be used to investigate fishery resources of land observers essential for the implementation of the TAC system. The system consists of an app and a cloud server that automatically measures the body size and height of fish and takes photos using a terminal equipped with a lidar sensor. In the cloud server, fish classification is performed using a CNN-based efficientnet model and the weight of fish is predicted using automatically measured body length and body height information. Using this system, it is possible to improve the existing method in which the land observer manually writes after measuring the tape measure and weight in the stomach market.

• Journal of Digital Convergence
• /
• v.12 no.8
• /
• pp.321-327
• /
• 2014

Feature selection is the technique to improve the classification performance by using a minimal set by removing features that are not related with each other and characterized by redundancy. This study proposed new feature selection using the distance between the center of gravity of the bounded sum of weighted fuzzy membership functions (BSWFMs) provided by the neural network with weighted fuzzy membership functions (NEWFM) in order to improve the classification performance. The distance-based feature selection selects the minimum features by removing the worst features with the shortest distance between the center of gravity of BSWFMs from the 24 initial features one by one, and then 22 minimum features are selected with the highest performance result. The proposed methodology shows that sensitivity, specificity, and accuracy are 97.7%, 99.7%, and 98.7% with 22 minimum features, respectively.

• Science of Emotion and Sensibility
• /
• v.21 no.1
• /
• pp.155-164
• /
• 2018

The temperature is a sense that can be felt by touch and sight. However, the concept of the temperature sensation is rarely used together with the concept of visual sensation and tactile sensation. In this study, the sensation of the temperature sensed through tactile and visual sense was investigated by the visual temperature depending on color and material characteristics. The textile was selected as a sample that could include color and material characteristics. The textile sample was composed of each 15-16 kinds of Yellow, Red, Blue, and Green of total 90 samples. The analytical method was to analyze first, the warm-cool of the colors of Yellow, Red, Blue, Green, and then to the visual temperature according to visual classification and tactile classification. And we investigated the correlation of the visual temperature depending on weight, thickness, and unevenness. As a result, the number of textiles felt by Cool and Warm differed according to the warm-cool of the colors feeling in the same textile. However, the visual temperature was different to each classification of textile. In particular, it was noticeable in thin, see-through and matte textiles. In relation to weight, thickness, unevenness and the visual temperature, the textile classification related to the weight is a classification of a hard, matte textile, and the textile classification related to the thickness is a thin, see-through textile.

• Korean Journal of Poultry Science
• /
• v.28 no.3
• /
• pp.187-192
• /
• 2001

The repeatabilities on Yolk percentage and yolk to albumen (Y:A) ratio of the eggs produced consecutively were investigated. The differences between two yolk lines in major egg characteristics were also evaluated. The investigations using one hundred ISA-Brown layers were conducted at 29 wk, 33 wk, 38 wk, and 43 wk of age. At the initiated age (29 wk of age) of the experiments, the birds producing eggs with lower or higher Y:A ratio than the overall mean Y:A ratio were classified as Low Yolk Line (LYL) or High Yolk Line (HYL), respectively. Overall, the eggs of LYL were significantly (P<0.05) lighter in yolk weight and lower in yolk percentage for the whole egg weight and Y:A ratio, but heavier in egg weight, albumen weight, and shell weight than those of HYL. The overall mean correlations among the three consecutive laying days in Y:A ratio showed highly significant (P<0.001) in both lines. At four different ages, the mean phenotypic correlation coefficients (r$_{p}$) among the three consecutive laying days in Y:A ratio also had very high significant(P<0.01 ~ 0.001) positive values. The egg weight was more closely associated with albumen weight than with yolk weight in both yolk lines.s.

• Proceedings of the Korean Fiber Society Conference
• /
• 1998.10a
• /
• pp.126-128
• /
• 1998

황토는 우리 나라의 주토이며 분해력, 자정력, 흡수력, 생명력을 지닌 적색토이다. 황토는 여러 가지 광물입자로 구성되어 있는데 그 크기는 0.02~0.05mm이며 다양한 크기의 입자들이 섞여 있으며 황토는 무게비로 50% 정도에 해당된다. 점토의 입자 크기는 0.005mm이하인 미세한 것을 말하며 약 5~10%정도 포함되어 있다. 따라서 황토를 광물염료로 분류할 수 있다. (중략)

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.1086-1087
• /
• 2023

본 논문은 물류창고 내 컨베이어벨트에서 자동으로 화물의 크기와 무게를 분석하고 이를 인공지능을 기반으로 분류하는 기술에 관한 연구를 다루고 있다. 우리의 연구를 통해 넓은 물류창고에서 전체 분류 과정을 모니터링할 수 있으며, 웹사이트를 활용하여 원거리에서도 물류 분류 과정을 실시간으로 확인 가능하게 한다. 또한 문제 발생 시 기록을 남겨 관리자 간에 관리, 감독이 원활하도록 도와준다.

• The Journal of the Acoustical Society of Korea
• /
• v.35 no.2
• /
• pp.83-91
• /
• 2016

Music-similarity computation is crucial in developing music information retrieval systems for browsing and classification. This paper overviews the recently-proposed centroid-model based music retrieval method and applies the distributional similarity measures to the model for retrieval-performance evaluation. Probabilistic distance measures (also called divergence) compute the distance between two probability distributions in a certain sense. In this paper, we consider the alpha divergence in computing distance between two centroid models for music retrieval. The alpha divergence includes the widely-used Kullback-Leibler divergence and Bhattacharyya distance depending on the values of alpha. Experiments were conducted on both genre and singer datasets. We compare the music-retrieval performance of the distributional similarity with that of the vector distances. The experimental results show that the alpha divergence improves the performance of the centroid-model based music retrieval.