• Journal of the Korean Institute of Intelligent Systems
• /
• v.25 no.5
• /
• pp.425-430
• /
• 2015

This paper is a study to improve the classification efficiency of rotating objects by using deep neural networks to which a deep learning algorithm was applied. For the classification experiment of rotating objects, COIL-20 is used as data and total 3 types of classifiers are compared and analyzed. 3 types of classifiers used in the study include PCA classifier to derive a feature value while reducing the dimension of data by using Principal Component Analysis and classify by using euclidean distance, MLP classifier of the way of reducing the error energy by using error back-propagation algorithm and finally, deep learning applied DBN classifier of the way of increasing the probability of observing learning data through pre-training and reducing the error energy through fine-tuning. In order to identify the structure-specific error rate of the deep neural networks, the experiment is carried out while changing the number of hidden layers and number of hidden neurons. The classifier using DBN showed the lowest error rate. Its structure of deep neural networks with 2 hidden layers showed a high recognition rate by moving parameters to a location helpful for recognition.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.37 no.7B
• /
• pp.586-594
• /
• 2012

The drastic growth of mobile communication and spreading of smart phone make the significant attention on Location Based Service. The one of most important things for vitalization of LBS is the accurate estimating position for mobile object. Focusing on AP's probabilistic position estimation, we develop an AP distribution map and new pattern matching algorithm for position estimation. The developed approaches can strengthen the advantages of Radio fingerprint based Wi-Fi Positioning System, especiall on the algorithms and data handling. Compared on the existing approaches of fingerprint pattern matching algorithm, we achieve the comparable higher performance on both of average error of estimation and deviation of errors. Furthermore all fingerprint data have been harvested from the actual measurement of radio fingerprint of Seoul, Kangnam area. This can approve the practical usefulness of proposed methodology.

• Transactions of the Korean Society of Automotive Engineers
• /
• v.8 no.2
• /
• pp.120-128
• /
• 2000

At the initial design stage of a new vehicle, the chassis layout has the most important influence on the overall vehicle performance. Most chassis designers have achieved the target performances by trial and error method as well as individual knowhow. Accordingly, a general procedure for determining the optimum location of suspension hard points with respect to the kinematic characteristics needs to be developed. In this paper, a method to optimize the toe angle in the double wishbone type front suspension of the four-wheel-drive vehicle is presented using the design of experiment, multibody dynamic simulation, and optimum design program. The handling performances of two full vehicle models having the initial and optimized toe angle are compared through the single lane change simulation. The sensitive design variables with respect to the kinematic characteristics are selected through the experimental design sensitivity analysis using the perturbation method. An object function is defined in terms of the toe angle among those kinematic characteristics. By the design of experiment and regression analysis, the regression model function of toe angle is obtained. The design variables which make the toe angle optimized ae extracted using the optimum design program DOT. The single lane change simulation and test of the full vehicle model are carried out to survey the handling performances of vehicle with toe angle optimized. The results of the single lane change simulation show that the optimized vehicle has the more improved understeer tendency than the initial vehicle.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2010.05a
• /
• pp.389-392
• /
• 2010

In this paper, it presents to detect location using structural information of hand from the input color images on Webcam and to recognize hand gestures. In this system, based on the skin color, the image changes a binary number and labels. Within each labeled area, we can find the Maximum Inscribed Circle using Voronoi Diagram. This circle can find the center of hand. And the circle extracts hand region from analyzing the ellipse elements to relate Maximum Inscribed Circle. We use the Maximum Inscribed Circle and the ellipse elements as characteristic of hand gesture recognition. In various environments, we cannot recognize the object that have similar colors like the background colors. But the proposed algorithm has the advantage that can be effectively eliminated about it.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.17 no.1
• /
• pp.67-75
• /
• 2024

Crosswalks play an important role for the safe movement of pedestrians in a complex urban environment. However, for the visually impaired, crosswalks can be a big risk factor. Although assistive tools such as braille blocks and acoustic traffic lights exist for safe walking, poor management can sometimes act as a hindrance to safety. This paper proposes a method to improve accuracy in a deep learning-based real-time crosswalk detection model that can be used in applications for pedestrian assistance for the disabled at the beginning. The image was binarized by utilizing the characteristic that the white line of the crosswalk image contrasts with the road surface, and through this, the crosswalk could be better recognized and the location of the crosswalk could be more accurately identified by using two models that learned the whole and the middle part of the crosswalk, respectively. In addition, it was intended to increase accuracy by creating a boundary box that recognizes crosswalks in two stages: whole and part. Through this method, additional frames that the detection model did not detect in RGB image learning from the crosswalk image could be detected.

• Journal of the Korean Society of Radiology
• /
• v.9 no.1
• /
• pp.17-21
• /
• 2015

Recently, As people's interest in the health of teeth is increased in the medical field changed into aging society, the number of times for the radiological diagnosis is increased. It can be said that the radiation exposure dose of Korean population is increased. It is also growing concern about radiation exposure. Therefore, the basic data for the dental panoramic X-ray system, its investigation and measuring the radiation dose is needed. In this study, we used ALOKA PDM-117 dosimeter and estimated a two-dimensional dose distribution of the dental panoramic X-ray system (VATEC Pax-400). Dose evaluation about the distribution is confirmed from the point of radiation exposure of a patient. Dose distribution of the dental panoramic X-ray system irradiated chin and the facial region to high dose as well as the parts of teeth. It was founded that the eye lens which are sensitive to radiation are exposed to unnecessary radiation, considering the effect of scattered radiation. The results of this study will be used more accurate dose assessment in a variety of object size and location of measuring dose.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.36 no.12C
• /
• pp.744-752
• /
• 2011

In intelligent surveillance system, various methods for detecting abnormal behavior were proposed recently. However, most researches are not robust enough to be utilized for actual reality which often has occlusions because of assumption the researches have that individual objects can be tracked. This paper presents a novel method to detect abnormal behavior by analysing major motion of the scene for complex environment in which object tracking cannot work. First, we generate Visual Word and Visual Document from motion information extracted from input video and process them through LDA(Latent Dirichlet Allocation) algorithm which is one of document analysis technique to obtain major motion information(location, magnitude, direction, distribution) of the scene. Using acquired information, we compare similarity between motion appeared in input video and analysed major motion in order to detect motions which does not match to major motions as abnormal behavior.

• Journal of KIISE:Databases
• /
• v.31 no.6
• /
• pp.624-640
• /
• 2004

Recently, with the rapid development of location-based techniques, index structures to efficiently manage moving objects have been required. In this paper, we propose a new spatio-temporal index structure that supports a future position retrieval and minimizes a update cost. The proposed index structure combines an assistant index structure that directly accesses current positions of moving objects with KDB-tree that is a space partitioning access method. The internal node in our proposed index structure keeps time parameters in order to support the future position retrieval and to minimize a update cost. Moreover, we propose new update and split methods to maximize the space utilization and the search performance. We perform various experiments to show that our proposed index structure outperforms the existing index structure.

• The Mathematical Education
• /
• v.60 no.3
• /
• pp.249-264
• /
• 2021

This article is devoted to investigating young learners' understanding of elapsed time from socio-cultural perspectives. The socio-cultural perspective benefits to access and personalize mathematics learning as how to have a mathematical object to be able to realize signifiers with the help of many other mathematical words and mediators. In terms of the realization of signifiers, I analyzed performances on elapsed time tasks of students in Grades 3 (n=115) and interviewed focal students. Quantitative analysis on students' performance identified that students perform differently when the task provided with the analog clock signifier. It suggested that students might think in a different way upon the given signifier even for the same elapsed time, especially when given as the analog clock. Qualitative analysis on focal students' interviews visualized how the students' understanding were different by displaying individual realization trees on elapsed time. The particular location of the analog clock signifier on each realization tree provided a personalized explanation about low performance on the task with analog clock signifier. The finding suggested that the realization of a specific signifier could be a key point in elapsed time understanding. I discussed why a majority of students face difficulty in elapsed time learning indicated analog clock and the advantage of moving elapsed time strands to higher grades in the school mathematics curriculum.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.24 no.3
• /
• pp.279-292
• /
• 2022

Unexpected anomalies in the geotechnical investigation at design stage may cause problems during tunnel excavation. Therefore, it is important to predict the ground condition ahead of a tunnel face during tunnel excavation in order to prevent tunnel collapse. Despite the fulfillment of an electrical resistivity survey at the tunnel face, the existing electrical resistivity survey program can produce distorted results by the limitation of tunnel modelling. In this background, this study develops a modelling program for an electrical resistivity survey considering the tunnel shape. Numerical simulation and inverse calculation were performed for the electrical resistivity survey in the tunnel using the developed program. As a result, it was proved that the developed program could predict accurately the anomalous object's location and condition ahead of the tunnel face.