• Proceedings of the Korea Information Processing Society Conference
• /
• 2012.04a
• /
• pp.556-559
• /
• 2012

CTC(Connected Target Coverage) 문제는 주어진 전체 타겟을 관측하고 관측한 데이터를 싱크노드까지 전송하는데 관여하는 센서집합의 개수를 최대화하여 네트워크 수명을 최대화하는 문제이다. 본 논문은 확률 센싱 및 연결성 모델을 기반으로 CTC문제에 접근한다. CTC문제를 해결하기 위해 휴리스틱 알고리즘인 CWGC-PM 알고리즘을 제안하고 시뮬레이션을 통해 알고리즘이 CTC문제를 해결하기에 적합함을 보인다. 또한 확률모델이 다양한 커버리지 및 연결성 요구조건에 적용될 수 있음을 보인다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2011.11a
• /
• pp.556-559
• /
• 2011

CTC(Connected Target Coverage)문제는 주어진 전체 타겟을 관측하고 관측한 데이터를 싱크노드까지 전송하는데 관여하는 센서집합의 개수를 최대화하여 네트워크 수명을 최대화하는 문제이다. 본 논문은 센서가 타겟을 관측할 확률이 타겟과의 거리에 영향을 받는다고 가정하는 확률 커버리지 모델을 기반으로 CTC문제에 접근한다. CTC문제를 해결하기 위해 휴리스틱 알고리즘인 CWGC-PM 알고리즘을 제안하고 시뮬레이션을 통해 알고리즘이 CTC문제를 해결하기에 적합함을 보인다.

• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.3
• /
• pp.115-122
• /
• 2019

For Korean phoneme recognition, Hidden Markov-Gaussian Mixture model(HMM-GMM) or hybrid models which combine artificial neural network with HMM have been mainly used. However, current approach has limitations in that such models require force-aligned corpus training data that is manually annotated by experts. Recently, researchers used neural network based phoneme recognition model which combines recurrent neural network(RNN)-based structure with connectionist temporal classification(CTC) algorithm to overcome the problem of obtaining manually annotated training data. Yet, in terms of implementation, these RNN-based models have another difficulty in that the amount of data gets larger as the structure gets more sophisticated. This problem of large data size is particularly problematic in the Korean language, which lacks refined corpora. In this study, we introduce CTC algorithm that does not require force-alignment to create a Korean phoneme recognition model. Specifically, the phoneme recognition model is based on convolutional neural network(CNN) which requires relatively small amount of data and can be trained faster when compared to RNN based models. We present the results from two different experiments and a resulting best performing phoneme recognition model which distinguishes 49 Korean phonemes. The best performing phoneme recognition model combines CNN with 3hop Bidirectional LSTM with the final Phoneme Error Rate(PER) at 3.26. The PER is a considerable improvement compared to existing Korean phoneme recognition models that report PER ranging from 10 to 12.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2018.05a
• /
• pp.398-401
• /
• 2018

본 연구에서는 심충 합성곱 신경망(Deep CNN)과 Connectionist Temporal Classification (CTC) 알고리즘을 사용하여 강제정렬 (force-alignment)이 이루어진 코퍼스 없이도 학습이 가능한 음소 인식 모델을 제안한다. 최근 해외에서는 순환 신경망(RNN)과 CTC 알고리즘을 사용한 딥 러닝 기반의 음소 인식 모델이 활발히 연구되고 있다. 하지만 한국어 음소 인식에는 HMM-GMM 이나 인공 신경망과 HMM 을 결합한 하이브리드 시스템이 주로 사용되어 왔으며, 이 방법 은 최근의 해외 연구 사례들보다 성능 개선의 여지가 적고 전문가가 제작한 강제정렬 코퍼스 없이는 학습이 불가능하다는 단점이 있다. 또한 RNN 은 학습 데이터가 많이 필요하고 학습이 까다롭다는 단점이 있어, 코퍼스가 부족하고 기반 연구가 활발하게 이루어지지 않은 한국어의 경우 사용에 제약이 있다. 이에 본 연구에서는 강제정렬 코퍼스를 필요로 하지 않는 CTC 알고리즘을 도입함과 동시에, RNN 에 비해 더 학습 속도가 빠르고 더 적은 데이터로도 학습이 가능한 합성곱 신경망(CNN)을 사용하여 딥 러닝 모델을 구축하여 한국어 음소 인식을 수행하여 보고자 하였다. 이 모델을 통해 본 연구에서는 한국어에 존재하는 49 가지의 음소를 추출하는 세 종류의 음소 인식기를 제작하였으며, 최종적으로 선정된 음소 인식 모델의 PER(phoneme Error Rate)은 9.44 로 나타났다. 선행 연구 사례와 간접적으로 비교하였을 때, 이 결과는 제안하는 모델이 기존 연구 사례와 대등하거나 조금 더 나은 성능을 보인다고 할 수 있다.

• Journal of Digital Contents Society
• /
• v.18 no.1
• /
• pp.93-99
• /
• 2017

A hybrid approach using Long Short Term Memory (LSTM) Recurrent Neural Network (RNN) has showed great improvement in speech recognition accuracy. For training acoustic model based on hybrid approach, it requires forced alignment of HMM state sequence from Gaussian Mixture Model (GMM)-Hidden Markov Model (HMM). However, high computation time for training GMM-HMM is required. This paper proposes an end-to-end approach for LSTM RNN-based Korean speech recognition to improve learning speed. A Connectionist Temporal Classification (CTC) algorithm is proposed to implement this approach. The proposed method showed almost equal performance in recognition rate, while the learning speed is 1.27 times faster.

• Journal of Broadcast Engineering
• /
• v.22 no.3
• /
• pp.381-396
• /
• 2017

In this paper, the perceptual rate control algorithm is studied for HEVC (High Efficiency Video Coding) encoder with bit allocation based on perceived visual quality. This paper proposes perceptual rate control algorithm which could consider perceived quality for HEVC encoding method. The proposed rate control algorithm employs adaptive bit allocation for frame and CTU level using the perceived visual importance of each CTU. For performance evaluation of the proposed algorithm, the proposed algorithm was implemented on HM 16.9 and tested for sequences in Class B under the CTC (Common Test Condition) RA (Random Access) case. Experimental results show that the proposed method reduces the bitrate of 3.12%, and improves BD-PSNR of 0.08dB and bitrate accuracy of 0.07% on average. And also, we achieved MOS improvement of 0.16 with the proposed method, compared with the conventional method based on DSCQS (Double Stimulus Continuous Quality Scale).

• Journal of Broadcast Engineering
• /
• v.21 no.6
• /
• pp.929-943
• /
• 2016

This paper proposes the rate control algorithm based on the S-JND (Saliency-Just Noticeable Difference) model for considering perceptual visual quality. The proposed rate control algorithm employs the S-JND model to simultaneously reflect human visual sensitivity and human visual attention for considering characteristics of human visual system. During allocating bits for CTU (Coding Tree Unit) level in a rate control, the bit allocation model calculates the S-JND threshold of each CTU in a picture. The threshold of each CTU is used for adaptively allocating a proper number of bits; thus, the proposed bit allocation model can improve perceptual visual quality. For performance evaluation of the proposed algorithm, the proposed algorithm was implemented on HM 16.9 and tested for sequences in Class B and Class C under the CTC (Common Test Condition) RA (Random Access), Low-delay B and Low-delay P case. Experimental results show that the proposed method reduces the bit-rate of 2.3%, and improves BD-PSNR of 0.07dB and bit-rate accuracy of 0.06% on average. We achieved MOS improvement of 0.03 with the proposed method, compared with the conventional method based on DSCQS (Double Stimulus Continuous Quality Scale).

• Geotechnical Engineering
• /
• v.3 no.1
• /
• pp.65-76
• /
• 1987

This study has been carried out as a basic course for the analysis of foundation deformations based on the Cap model using the finite element methods. Material parameters should firstly be determined in order to use the Cap model for numerical solution. Associated with the fact described above, a method determining the soil parameters is suggested using algorithm for numerical ana])isis from raw truly triaxial compression laboratory test data of Pueblo.Colorado sand by Zaman, et at. (1982) More specifically, the change of soil parameters Is thoroughly examined by weighting the data obtained from CTC and RTE tests, respectively. The main results obtained are as follows; 1. The obtained values of parameters (E, V and 2) are same irrespective of data obtained from various kind of tests. 2. The values of the other parameters are dependent on data used. 3. The determination of parameters is little affected by the weighting factor.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.35 no.3B
• /
• pp.453-461
• /
• 2010

A critical issue in wireless sensor networks is an energy-efficiency since the sensor batteries have limited energy power and, in most cases, are not rechargeable. The most practical manner relate to this issue is to use a node wake-up scheduling protocol that some sensor nodes stay active to provide sensing service, while the others are inactive for conserving their energy. Especially, CTC (Connected Target Coverage) problem has been considered as a representative energy-efficiency problem considering connectivity as well as target coverage. In this paper, we propose a new energy consumption model considering multiple-targets and create a new problem, CMTC (Connected Multiple-Target Coverage) problem, of which objective is to maximize the network lifetime based on the energy consumption model. Also, we present SPT (Shortest Path based on Targets)-Greedy algorithm to solve the problem. Our simulation results show that SPT-Greedy algorithm performs much better than previous algorithm in terms of the network lifetime.

• Proceedings of the KSR Conference
• /
• 2005.05a
• /
• pp.1159-1164
• /
• 2005

현재 철도청이 운용중인 열차운행관리 시스템(Railway Traffic Management System, RTMS)은 서울, 대전, 부산, 순천 그리고 영주 등으로 총 5개 지역본부로 분산되어 있어 업무의 중복을 줄이고, 자동화(Automation)된 열차집중제어장치(Central Traffic Control, CTC)를 구축하기 위해 지역본부를 대전으로 통합하는 프로젝트를 진행중이다. 본 논문은 철도청 사령실 통합 신호설비 구축 프로젝트에 의거하여 열차 경합을 검지 또는 예측하고 운영자에게 최소의 시간 내에 최적의 해소 대책을 제시함을 목적으로 하는 열차 경합 검지 시스템을 구현하는 과정에서 작성되었다. 여기에서는 열차 경합 검지에 대한 개요와 검지 가능한 경합 종류에 대해 기술하고, 실제 구현된 알고리즘의 기본적인 내용, 프로세스의 구성도 및 시뮬레이션 결과를 설명하려고 한다.