• Journal of Institute of Control, Robotics and Systems
• /
• v.15 no.11
• /
• pp.1115-1121
• /
• 2009

For HRI (Human-Robot Interaction) in daily life, robots need to recognize non-rigid objects such as clothes and blankets. However, the recognition of non-rigid objects is challenging because of the variation of the shapes according to the places and laying manners. In this paper, the cognition of non-rigid object based on a cognitive system is presented. The characteristics of non-rigid objects are analysed in the view of HRI and referred to design a framework for the cognition of them. We adopt a linguistic cognitive system for describing all of the events happened to robots. When an event related to the non-rigid objects is occurred, the cognitive system describes the event into a sentential form and stores it at a sentential memory, and depicts the objects with a spatial model for being used as references. The cognitive system parses each sentence syntactically and semantically, in which the nouns meaning objects are connected to their models. For answering the questions of humans, sentences are retrieved by searching temporal information in the sentential memory and by spatial reasoning in a schematic imagery. Experiments show the feasibility of the cognitive system for cognizing non-rigid objects in HRI.

• Biomedical Science Letters
• /
• v.26 no.1
• /
• pp.14-21
• /
• 2020

Quantitative evaluation of the sense of smell is done by the olfactory event-related potential (OERP) test. OERP consists of N1, P1, N2, P2, and P3, of which N1 and P1 latency are known to be affected by the concentration of odor stimuli associated with the pre-sensory phase and P2 and P3 are the stages at which odors are perceived and are known to change by subjective evaluation of the stimulus. The purpose of this study was to clarify the correlation between the expression of OERP and the subjective evaluation of the object on the fragrance stimulus using various fragrances. Therefore, the study examined the relative waveform power ratio, preference for each scent stimulus, and finally the amplitude and latency change of the components of OERP, N1 and P2. In contrast, it was found that the late P2 response waveform was an effective aroma stimulus recognition waveform in OERP compared to the initial response waveforms.

• Proceedings of the ESK Conference
• /
• 1996.04a
• /
• pp.113-127
• /
• 1996

In Human Reliability Analysis(HRA), the uncertainties involved in many factors that affect human reliability have to be represented as the quantitative forms. Conventional probability- based human reliability theory is used to evaluate the effect of those uncertainties but it is pointed out that the actual human reliability should be different from that of conventional one. Conventional HRA makes use of error rates, however, it is difficult to collect data enough to estimate these error rates, and the estimates of error rates are dependent only on engineering judgement. In this paper, the error possibility that is proposed by Onisawa is used to represent human reliability, and the error possibility is obtained by use of fuzzy reasoning that plays an important role to clarify the relation between human reliability and human error. Also, assuming these factors are connected to the top event through Fault Tree structure, the influence and correlation of these factors are measured by fuzzy operation. When a fuzzy operation is applied to Fault Tree Analysis, it is possible to simplify the operation applying the logic disjuction and logic conjuction to structure function, and the structure of human reliability can be represented as membership function of the top event. Also, on the basis of the the membership function, the characteristics of human reliability can be evaluated by use of the concept of pattern recognition.

• KIPS Transactions on Software and Data Engineering
• /
• v.9 no.6
• /
• pp.177-186
• /
• 2020

Recent advances in deep learning technology have improved image recognition performance in the field of computer vision, and serverless computing is emerging as the next generation cloud computing technology for event-based cloud application development and services. Attempts to use deep learning and serverless computing technology to increase the number of real-world image recognition services are increasing. Therefore, this paper describes how to develop an efficient deep learning based image recognition service system using serverless computing technology. The proposed system suggests a method that can serve large neural network model to users at low cost by using AWS Lambda Server based on serverless computing. We also show that we can effectively build a serverless computing system that uses a large neural network model by addressing the shortcomings of AWS Lambda Server, cold start time and capacity limitation. Through experiments, we confirmed that the proposed system, using AWS Lambda Serverless Computing technology, is efficient for servicing large neural network models by solving processing time and capacity limitations as well as cost reduction.

• The Journal of the Acoustical Society of Korea
• /
• v.38 no.6
• /
• pp.670-677
• /
• 2019

As the presence of background noise in acoustic signal degrades the performance of speech or acoustic event recognition, it is still challenging to extract noise-robust acoustic features from noisy signal. In this paper, we propose a combined structure of Wasserstein Generative Adversarial Network (WGAN) and MultiTask AutoEncoder (MTAE) as deep learning architecture that integrates the strength of MTAE and WGAN respectively such that it estimates not only noise but also speech features from noisy acoustic source. The proposed MTAE-WGAN structure is used to estimate speech signal and the residual noise by employing a gradient penalty and a weight initialization method for Leaky Rectified Linear Unit (LReLU) and Parametric ReLU (PReLU). The proposed MTAE-WGAN structure with the adopted gradient penalty loss function enhances the speech features and subsequently achieve substantial Phoneme Error Rate (PER) improvements over the stand-alone Deep Denoising Autoencoder (DDAE), MTAE, Redundant Convolutional Encoder-Decoder (R-CED) and Recurrent MTAE (RMTAE) models for robust speech recognition.

• The Journal of Korean Institute of Next Generation Computing
• /
• v.14 no.5
• /
• pp.53-59
• /
• 2018

Due to the recent developments in computer vision technology, complex actions can be recognized with reasonable accuracy in smart cities. In contrast, violence recognition such as events related to fight and knife, has gained less attention. The capability of visual surveillance can be used for detecting fight in streets or in prison centers. In this paper, we proposed a deep learning-based violence recognition method for surveillance cameras. A convolutional neural network (CNN) model is trained and fine-tuned on available benchmark datasets of fights and knives for violence recognition. When an abnormal event is detected, an alarm can be sent to the nearest police station to take immediate action. Moreover, when the probabilities of fight and knife classes are predicted very low, this situation is considered as normal situation. The experimental results of the proposed method outperformed other state-of-the-art CNN models with high margin by achieving maximum 99.21% accuracy.

• International conference on construction engineering and project management
• /
• 2020.12a
• /
• pp.313-322
• /
• 2020

The recognition of the risk hazards is a vital step to effectively prevent accidents on a construction site. The advanced development in computer vision systems and the availability of the large visual database related to construction site made it possible to take quick action in the event of human error and disaster situations that may occur during management supervision. Therefore, it is necessary to analyze the risk factors that need to be managed at the construction site and review appropriate and effective technical methods for each risk factor. This research focuses on analyzing Occupational Safety and Health Agency (OSHA) related to risk zone identification rules that can be adopted by the image recognition technology and classify their risk factors depending on the effective technical method. Therefore, this research developed a pattern-oriented classification of OSHA rules that can employ a large scale of safety hazard recognition. This research uses joint reasoning of risk zone Identification and numeric input by utilizing a stereo camera integrated with an image detection algorithm such as (YOLOv3) and Pyramid Stereo Matching Network (PSMNet). The research result identifies risk zones and raises alarm if a target object enters this zone. It also determines numerical information of a target, which recognizes the length, spacing, and angle of the target. Applying image detection joint logic algorithms might leverage the speed and accuracy of hazard detection due to merging more than one factor to prevent accidents in the job site.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.10 no.4
• /
• pp.189-198
• /
• 2015

Most IoT-related approaches have tried to establish the relation by connecting the network between things. The proposed research will present how the pervasive interaction of eco-system formed by touching the objects between humans and things can be recognized on purpose. By collecting and sharing the detected patterns among all kinds of things, we can construct the environment which enables individualized interactions of different objects. To perform the aforementioned, we are going to utilize technical procedures such as event-driven signal processing, pattern matching for signal recognition, and hardware in the loop simulation. We will also aim to implement the prototype of sensor processor based on Arduino MCU, which can be integrated with system using Arduino-Matlab/Simulink hybrid-interoperation environment. In the experiment, we use piezo transducer to detect the vibration or vibrates the surface using acoustic wave, which has specific frequency spectrum and individualized signal shape in terms of time axis. The signal distortion in time and frequency domain is recorded into memory tracer within sensor processor to extract the meaningful pattern by comparing the stored with lookup table(LUT). In this paper, we will contribute the initial prototypes for the acoustic touch processor by using off-the-shelf MCU and the integrated framework based on Matlab/Simulink model to provide the individualization of the touch-sensing for the user on purpose.

• Journal of Global Scholars of Marketing Science
• /
• v.15 no.2
• /
• pp.123-140
• /
• 2005

Recently, one of the issues that are significantly considered is to keep Changwon F3 car racing. In most existing studies, however, there is shortage of opinions for understanding the reactions of the host community. The question addressed in this paper is what are the underlying factors that affect which inhibitors are likely to be discussed by the host community for car racing. Survey collected from 299 participants in Changwon city were analyzed to test their reactions. Through the empirical study, several research findings emerged. The results indicate that (1) the event seeker, the drive seeker, and the hard internal seeker were significant discriminant factors to separate inhibitors from the host community, (2) the significant differences between the recognition of demographic characteristics and that of negative effects, (3) the significant level between cluster groups(overcontrolled, resilient, undercontrolled) and negative effects is found. The findings also provide implications for practice on several fronts, which is to understand the reaction of the host community to F3 car racing, and for the event planner to reduce complaints.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.53 no.12
• /
• pp.152-158
• /
• 2016

In this paper, we implemented the full stack platform design with MongoDB database of open source platform Raspberry PI 3 model. We experimented the triggering of event driven with acceleration sensor data logging with wireless communication. we captured the image of USB Camera(MS LifeCam cinema) with 28 frames per second under the Linux version of Raspbian Jessie and extended the functionality of wireless communication function with Bluetooth technology for the purpose of making Android Mobile devices interface. And therefore we implemented the functions of the full stack platform for recognizing the event triggering characteristics of detecting the acceleration sensor action and gathering the temperature and humidity sensor data under IoT environment. Especially we used MEAN Stack for developing the performance of full stack platform because the MEAN Stack is more akin to working with MongoDB than what we know of as a database. Afterwards, we would enhance the performance of full stack platform for IoT clouding functionalities and more feasible web design with MongoDB.