• The KIPS Transactions:PartB
• /
• v.11B no.7 s.96
• /
• pp.885-892
• /
• 2004

In this paper, we propose a multi-modal biometrics system based on face and signature recognition system. Here, the face recognition system is designed by fuzzy LDA, and the signature recognition system is implemented with the LDA and segment matching methods. To effectively aggregate two systems, we obtain statistical distribution models based on matching values for genuine and impostor, respectively. And then, the final verification is Performed by the support vector machine. From the various experiments, we find that the proposed method shows high recognition rates comparing with the conventional methods.

• Journal of Distribution Science
• /
• v.17 no.8
• /
• pp.5-14
• /
• 2019

Purpose - This study aims to proposes a conceptual framework to segment multi-national products based on a Chinese consumer's perception of multi-national products, to find the role of consumer ethnocentrism (CET) in country of origin (COO) effects for Chinese, and to figure out how different dimension of CET Effects on purchase intention developed market and home country. Research design, data and methodology - This study selected a 2×2×2 factorial design for the hypothesis test based on the product category × combination of manufactured type × Ethnocentrism level. This study distinguishes products between luxury (Burberry) and non-luxury (Nike) products and choose combination of manufactured type (Spain vs India/ Spain vs China) in order to perform comparative studies. A total of 223 Chinese participated in the experiment. After being exposed to each scenario, participants were asked to respond to questions about brand preference and purchase intention Results - Regarding to luxury made in developed country, it is worth that exposing COO information to low level of ethnocentrism consumers. Regarding to non-luxury product made in emerging country, it makes it worse when COO information to high level of ethnocentrism consumers. Lastly, regarding to non-luxury product, patriotic consumers prefer to purchase product made in home country.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.10 no.2
• /
• pp.647-669
• /
• 2016

Multiple attribute for decision making including user preference will increase the complexity of route selection process. Various approaches have been proposed to solve the optimal route selection problem. In this paper, multi attribute decision making (MADM) algorithms such as Simple Additive Weighting (SAW), Weighted Product Method (WPM), Analytic Hierarchy Process (AHP) method and Total Order Preference by Similarity to the Ideal Solution (TOPSIS) methods have been proposed for acoustic signature based optimal route selection to facilitate user with better quality of service. The traffic density state conditions (very low, low, below medium, medium, above medium, high and very high) on the road segment is the occurrence and mixture weightings of traffic noise signals (Tyre, Engine, Air Turbulence, Exhaust, and Honks etc) is considered as one of the attribute in decision making process. The short-term spectral envelope features of the cumulative acoustic signals are extracted using Mel-Frequency Cepstral Coefficients (MFCC) and Adaptive Neuro-Fuzzy Classifier (ANFC) is used to model seven traffic density states. Simple point method and AHP has been used for calculation of weights of decision parameters. Numerical results show that WPM, AHP and TOPSIS provide similar performance.

• The Journal of the Acoustical Society of Korea
• /
• v.26 no.2E
• /
• pp.39-43
• /
• 2007

Music summarization refers to a technique which automatically extracts the most important and representative segments in music content. In this paper, we propose and evaluate a technique which provides the repeated part in music content as music summary. For extracting a repeated segment in music content, the proposed algorithm uses the weighted sum of similarity measures based on multi-level vector quantization for fixed-length summary or optimal-length summary. For similarity measures, count-based similarity measure and distance-based similarity measure are proposed. The number of the same codeword and the Mahalanobis distance of features which have same codeword at the same position in segments are used for count-based and distance-based similarity measure, respectively. Fixed-length music summary is evaluated by measuring the overlapping ratio between hand-made repeated parts and automatically generated ones. Optimal-length music summary is evaluated by calculating how much automatically generated music summary includes repeated parts of the music content. From experiments we observed that optimal-length summary could capture the repeated parts in music content more effectively in terms of summary length than fixed-length summary.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.7 no.3
• /
• pp.198-204
• /
• 2007

In mobile robotics, ultrasonic sensors became standard devices for collision avoiding. Moreover, their applicability for map building and navigation has exploited in recent years. In this paper, as the preliminary step for developing a multi-purpose autonomous carrier mobile robot to transport trolleys or heavy goods and serve as robotic nursing assistant in hospital wards. The aim of this paper is to present the use of multi-sensor data fusion such as ultrasonic sensor, IR sensor for mobile robot to navigate, and presents an experimental mobile robot designed to operate autonomously within both indoor and outdoor environments. The global map building based on multi-sensor data fusion is applied for recognition an obstacle free path from a starting position to a known goal region, and simultaneously build a map of straight line segment geometric primitives based on the application of the Hough transform from the actual and noisy sonar data. We will give an explanation for the robot system architecture designed and implemented in this study and a short review of existing techniques, Hough transform, since there exist several recent thorough books and review paper on this paper. Experimental results with a real Pioneer DX2 mobile robot will demonstrate the effectiveness of the discussed methods.

• Journal of Korea Multimedia Society
• /
• v.24 no.8
• /
• pp.1000-1011
• /
• 2021

In recent years, using Deep Learning methods to apply for medical and biomedical image analysis has seen many advancements. In clinical, using Deep Learning-based approaches for cancer image analysis is one of the key applications for cancer detection and treatment. However, the scarcity and shortage of labeling images make the task of cancer detection and analysis difficult to reach high accuracy. In 2015, the Unet model was introduced and gained much attention from researchers in the field. The success of Unet model is the ability to produce high accuracy with very few input images. Since the development of Unet, there are many variants and modifications of Unet related architecture. This paper proposes a new approach of using Unet++ with pretrained EfficientNet as backbone architecture for breast tumor cell nuclei segmentation and uses the multi-organ transfer learning approach to segment nuclei of breast tumor cells. We attempt to experiment and evaluate the performance of the network on the MonuSeg training dataset and Triple Negative Breast Cancer (TNBC) testing dataset, both are Hematoxylin and Eosin (H & E)-stained images. The results have shown that EfficientUnet++ architecture and the multi-organ transfer learning approach had outperformed other techniques and produced notable accuracy for breast tumor cell nuclei segmentation.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.10
• /
• pp.3230-3255
• /
• 2022

Causality mining in NLP is a significant area of interest, which benefits in many daily life applications, including decision making, business risk management, question answering, future event prediction, scenario generation, and information retrieval. Mining those causalities was a challenging and open problem for the prior non-statistical and statistical techniques using web sources that required hand-crafted linguistics patterns for feature engineering, which were subject to domain knowledge and required much human effort. Those studies overlooked implicit, ambiguous, and heterogeneous causality and focused on explicit causality mining. In contrast to statistical and non-statistical approaches, we present Bidirectional Encoder Representations from Transformers (BERT) integrated with Multi-level Feature Networks (MFN) for causality recognition, called BERT+MFN for causality recognition in noisy and informal web datasets without human-designed features. In our model, MFN consists of a three-column knowledge-oriented network (TC-KN), bi-LSTM, and Relation Network (RN) that mine causality information at the segment level. BERT captures semantic features at the word level. We perform experiments on Alternative Lexicalization (AltLexes) datasets. The experimental outcomes show that our model outperforms baseline causality and text mining techniques.

• Journal of Platform Technology
• /
• v.9 no.2
• /
• pp.38-45
• /
• 2021

A frequently used bioimpedance analytical method in Korea is the segmental multi-frequency BIA (SMF-BIA) method, but it is not directly determined at a segmented impedance. This study was to compare SMF-BIA determinations with direct segmented determinations for accuracy and appropriateness of segment parameters. This study is to compare the segment parameters, accuracy and appropriateness of the multi-frequency segmental bioimpedance analysis. To this end, 108 elderly individuals were measured. Segmented bioelectrical measurements obtained from a SMF-BIA (Inbody S10) at 50 kHz and measured with a phase sensitive single frequency device (SF-BIA, bia-101, RJL / akern systems) were compared. The significant difference (%) was demonstrated between single - and multiple frequency determinations of the right upper limb (R = 35.5 ± 6.2%, P < 0.001; Xc = 2.7 ± 7.6%, P < 0.01), left upper limb difference (R= 33. 9 ± 6.0%, P < 0.001; Xc = 2.8 ± 8.3%, P < 0.01), right lower limb difference (R = 18.6 ± 4.3%, P < 0.001; Xc = 25.8 ± 10.0%, P < 0.001), left lower limb difference (R = 18.0 ± 4.7%, P < 0.001; Xc = 31.8%). Of the results determined with the two BIA methods, the impedance measurements of the limbs and whole body showed a high correlation (RA: R = 0. 950, LA: R = 0. 949, RL: R = 0.899, LL: R = 0.88), and in the agreement test, the impedance values of the upper limbs and whole body also showed strong agreement (ICC > 0.9), but in the Xc, the correlation was weak. In conclusion, it was found that although bioimpedance devices had significantly different characteristics and inconsistent cross sectionally, there was a high population level agreement in the upper and lower extremities in determining segmental resistance value changes. But a large error was found on the trunk. Further studies were needed for reducing the error.

• Korean Journal of Applied Biomechanics
• /
• v.15 no.3
• /
• pp.143-152
• /
• 2005

The purpose of this study was to predict the failure or success of the Snatch-lifting trial as a consequence of the stand-up phase simulated in Kane's equation of motion that was effective for the dynamic analysis of multi-segment. This experiment was a case study in which one male athlete (age: 23yrs, height: 154.4cm, weight: 64.5kg) from K University was selected The system of a simulation included a multi-segment system that had one degree of freedom and one generalized coordinate for the shank segment angle. The reference frame was fixed by the Nonlinear Trans formation (NLT) method in order to set up a fixed Cartesian coordinate system in space. A weightlifter lifted a 90kg-barbell that was 75% of subject's maximum lifting capability (120kg). For this study, six cameras (Qualisys Proreflex MCU240s) and two force-plates (Kistler 9286AAs) were used for collecting data. The motion tracks of 11 land markers were attached on the major joints of the body and barbell. The sampling rates of cameras and force-plates were set up 100Hz and 1000Hz, respectively. Data were processed via the Qualisys Track manager (QTM) software. Landmark positions and force-plate amplitudes were simultaneously integrated by Qualisys system The coordinate data were filtered using a fourth-order Butterworth low pass filtering with an estimated optimum cut-off frequency of 9Hz calculated with Andrew & Yu's formula. The input data of the model were derived from experimental data processed in Matlab6.5 and the solution of a model made in Kane's method was solved in Matematica5.0. The conclusions were as follows; 1. The torque motor of the shank with 246Nm from this experiment could lift a maximum barbell weight (158.98kg) which was about 246 times as much as subject's body weight (64.5kg). 2. The torque motor with 166.5 Nm, simulated by angular displacement of the shank matched to the experimental result, could lift a maximum barbell weight (90kg) which was about 1.4 times as much as subject's body weight (64.5kg). 3. Comparing subject's maximum barbell weight (120kg) with a modeling maximum barbell weight (155.51kg) and with an experimental maximum barbell weight (90kg), the differences between these were about +35.7kg and -30kg. These results strongly suggest that if the maximum barbell weight is decided, coaches will be able to provide further knowledge and information to weightlifters for the performance improvement and then prevent injuries from training of weightlifters. It hopes to apply Kane's method to other sports skill as well as weightlifting to simulate its motion in the future study.

• KIPS Transactions on Computer and Communication Systems
• /
• v.9 no.2
• /
• pp.37-44
• /
• 2020

In this paper, we proposed that a fully automatic multi-class whole heart segmentation algorithm using deep learning. The proposed method is based on U-Net architecture which consist of recurrent convolutional block, residual multi-dilated convolutional block. The evaluation was accomplished by comparing automated analysis results of the test dataset to the manual assessment. We obtained the average DSC of 96.88%, precision of 95.60%, and recall of 97.00% with CT images. We were able to observe and analyze after visualizing segmented images using three-dimensional volume rendering method. Our experiment results show that proposed method effectively performed to segment in various heart structures. We expected that our method can help doctors and radiologist to make image reading and clinical decision.