• Smart Structures and Systems
• /
• 제8권1호
• /
• pp.69-92
• /
• 2011

This paper presents an emergent pattern recognition approach based on the immune network theory and hierarchical clustering algorithms. The immune network allows its components to change and learn patterns by changing the strength of connections between individual components. The presented immune-network-based approach achieves emergent pattern recognition by dynamically generating an internal image for the input data patterns. The members (feature vectors for each data pattern) of the internal image are produced by an immune network model to form a network of antibody memory cells. To classify antibody memory cells to different data patterns, hierarchical clustering algorithms are used to create an antibody memory cell clustering. In addition, evaluation graphs and L method are used to determine the best number of clusters for the antibody memory cell clustering. The presented immune-network-based emergent pattern recognition (INEPR) algorithm can automatically generate an internal image mapping to the input data patterns without the need of specifying the number of patterns in advance. The INEPR algorithm has been tested using a benchmark civil structure. The test results show that the INEPR algorithm is able to recognize new structural damage patterns.

• 한국음향학회지
• /
• 제18권5호
• /
• pp.44-51
• /
• 1999

본 논문에서는 음성과 자연언어의 통합처리를 위한 효과적인 병렬계산모델을 제안한다. 음소모델은 연속 Hidden Markov Model(HMM)에 기반을 둔 문맥종속형 음소를 사용하며, 언어모델은 지식베이스를 기반으로 한다. 또한 지식베이스를 구성하기 위해 계층구조의 semantic network과 병렬 marker-passing을 추론 메카니즘으로 쓰는 memory-based parsing 기술을 사용한다. 본 연구의 병렬 음성인식 알고리즘은 분산메모리 MIMD(Multiple Instruction Multiple Data) 구조의 다중 Transputer 시스템을 이용하여 구현되었다. 실험결과, 본 연구의 지식베이스 기반 음성인식 시스템의 인식률이 word network 기반 음성인식 시스템보다 높게 나타났으며 code-phoneme 통계정보를 활용하여 인식성능의 향상도 얻을 수 있었다. 또한, 성능향상도(speedup) 관련 실험들을 통하여 병렬 음성인식 시스템의 실시간 구현 가능성을 확인하였다.

• 인터넷정보학회논문지
• /
• 제21권5호
• /
• pp.67-73
• /
• 2020

In this paper, we present a CNN-based gesture recognition approach which reduces the memory burden of input data. Most of the neural network-based gesture recognition methods have used a sequence of frame images as input data, which cause a memory burden problem. We use a motion history image in order to define a meaningful gesture. The motion history image is a grayscale image into which the temporal motion information is collapsed by synthesizing silhouette images of a user during the period of one meaningful gesture. In this paper, we first summarize the previous traditional approaches and neural network-based approaches for gesture recognition. Then we explain the data preprocessing procedure for making the motion history image and the neural network architecture with three convolution layers for recognizing the meaningful gestures. In the experiments, we trained five types of gestures, namely those for charging power, shooting left, shooting right, kicking left, and kicking right. The accuracy of gesture recognition was measured by adjusting the number of filters in each layer in the proposed network. We use a grayscale image with 240 × 320 resolution which defines one meaningful gesture and achieved a gesture recognition accuracy of 98.24%.

• 한국컴퓨터정보학회논문지
• /
• 제10권4호
• /
• pp.17-27
• /
• 2005

본 논문에서는 문자인식과 복원을 위한 신경회로망 시스템을 제안한다. 제안하는 시스템은 인식부와 연상부로 구성되었다. 인식부에서는 ART 신경회로망의 인식성능을 개선하기 위해 불필요한 하향틀의 생성과 변화를 제한하여 효과적인 패턴인식이 가능한 모델을 제안하였다. 또한, 한글의 구조적인 특징을 능동적으로 적용할 수 있게 구성된 위치특징 추출 알고리즘을 적용하였다. 연상부에서는 Hopfield 신경회로망으로, 입력된 이미지 패턴의 복원이 가능한 모델을 구성하였다. 제안하는 시스템은 그 성능을 확인하기 위해 각 부분별 실험을 하였다. 그 결과 인식율이 개선되고 복원이 가능함을 보였다.

• 대한음성학회지:말소리
• /
• 제50호
• /
• pp.51-69
• /
• 2004

In this paper, we present a one-pass semi-dynamic network decoding framework that inherits both advantages of fast decoding speed from static network decoders and memory efficiency from dynamic network decoders. Our method is based on the novel language model network representation that is essentially of finite state machine (FSM). The static network derived from the language model network [1][2] is partitioned into smaller subnetworks which are static by nature or self-structured. The whole network is dynamically managed so that those subnetworks required for decoding are cached in memory. The network is near-minimized by applying the tail-sharing algorithm. Our decoder is evaluated on the 25k-word Korean broadcast news transcription task. In case of the search network itself, the network is reduced by 73.4% from the tail-sharing algorithm. Compared with the equivalent static network decoder, the semi-dynamic network decoder has increased at most 6% in decoding time while it can be flexibly adapted to the various memory configurations, giving the minimal usage of 37.6% of the complete network size.

• ETRI Journal
• /
• 제44권4호
• /
• pp.672-685
• /
• 2022

This article proposes a subject-independent application of brain-computer interfacing (BCI). A 32-channel Electroencephalography (EEG) device is used to measure imagined speech (SI) of four words (sos, stop, medicine, washroom) and one phrase (come-here) across 13 subjects. A deep long short-term memory (LSTM) network has been adopted to recognize the above signals in seven EEG frequency bands individually in nine major regions of the brain. The results show a maximum accuracy of 73.56% and a network prediction time (NPT) of 0.14 s which are superior to other state-of-the-art techniques in the literature. Our analysis reveals that the alpha band can recognize SI better than other EEG frequencies. To reinforce our findings, the above work has been compared by models based on the gated recurrent unit (GRU), convolutional neural network (CNN), and six conventional classifiers. The results show that the LSTM model has 46.86% more average accuracy in the alpha band and 74.54% less average NPT than CNN. The maximum accuracy of GRU was 8.34% less than the LSTM network. Deep networks performed better than traditional classifiers.

• 한국멀티미디어학회논문지
• /
• 제23권8호
• /
• pp.1019-1029
• /
• 2020

In this paper, we propose a novel knowledge distillation algorithm to create an compressed deep ensemble network coupled with the combined use of local and global features of face images. In order to transfer the capability of high-level recognition performances of the ensemble deep networks to a single deep network, the probability for class prediction, which is the softmax output of the ensemble network, is used as soft target for training a single deep network. By applying the knowledge distillation algorithm, the local feature informations obtained by training the deep ensemble network using facial subregions of the face image as input are transmitted to a single deep network to create a so-called compressed ensemble DCNN. The experimental results demonstrate that our proposed compressed ensemble deep network can maintain the recognition performance of the complex ensemble deep networks and is superior to the recognition performance of a single deep network. In addition, our proposed method can significantly reduce the storage(memory) space and execution time, compared to the conventional ensemble deep networks developed for face recognition.

• 한국음향학회지
• /
• 제38권6호
• /
• pp.703-709
• /
• 2019

본 논문에서는 구개인두부전증(VeloPharyngeal Insufficiency, VPI) 환자의 음성을 효과적으로 인식하기 위해 컨볼루션 신경망 (Convolutional Neural Network, CNN), 장단기 모델(Long Short Term Memory, LSTM) 구조 신경망을 은닉 마르코프 모델(Hidden Markov Model, HMM)과 결합한 하이브리드 구조의 음성 인식 시스템을 구축하고 모델 적응 기법을 적용하여, 기존 Gaussian Mixture Model(GMM-HMM), 완전 연결형 Deep Neural Network(DNN-HMM) 기반의 음성 인식 시스템과 성능을 비교한다. 정상인 화자가 PBW452단어를 발화한 데이터를 이용하여 초기 모델을 학습하고 정상인 화자의 VPI 모의 음성을 이용하여 화자 적응의 사전 모델을 생성한 후에 VPI 환자들의 음성으로 추가 적응 학습을 진행한다. VPI환자의 화자 적응 시에 CNN-HMM 기반 모델에서는 일부층만 적응 학습하고, LSTM-HMM 기반 모델의 경우에는 드롭 아웃 규제기법을 적용하여 성능을 관찰한 결과 기존 완전 연결형 DNN-HMM 인식기보다 3.68 % 향상된 음성 인식 성능을 나타낸다. 이러한 결과는 본 논문에서 제안하는 LSTM-HMM 기반의 하이브리드 음성 인식 기법이 많은 데이터를 확보하기 어려운 VPI 환자 음성에 대해 보다 향상된 인식률의 음성 인식 시스템을 구축하는데 효과적임을 입증한다.

• 디지털콘텐츠학회 논문지
• /
• 제18권1호
• /
• pp.93-99
• /
• 2017

Long Short Term Memory (LSTM) Recurrent Neural Network (RNN)를 이용한 hybrid 방법은 음성 인식률을 크게 향상시켰다. Hybrid 방법에 기반한 음향모델을 학습하기 위해서는 Gaussian Mixture Model (GMM)-Hidden Markov Model (HMM)로부터 forced align된 HMM state sequence가 필요하다. 그러나, GMM-HMM을 학습하기 위해서 많은 연산 시간이 요구되고 있다. 본 논문에서는 학습 속도를 향상하기 위해, LSTM RNN 기반 한국어 음성인식을 위한 end-to-end 방법을 제안한다. 이를 구현하기 위해, Connectionist Temporal Classification (CTC) 알고리즘을 제안한다. 제안하는 방법은 기존의 방법과 비슷한 인식률을 보였지만, 학습 속도는 1.27 배 더 빨라진 성능을 보였다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제13권11호
• /
• pp.5580-5593
• /
• 2019

A malicious Uniform Resource Locator (URL) recognition and detection method based on the combination of Attention mechanism with Convolutional Neural Network and Long Short-Term Memory Network (Attention-Based CNN-LSTM), is proposed. Firstly, the WHOIS check method is used to extract and filter features, including the URL texture information, the URL string statistical information of attributes and the WHOIS information, and the features are subsequently encoded and pre-processed followed by inputting them to the constructed Convolutional Neural Network (CNN) convolution layer to extract local features. Secondly, in accordance with the weights from the Attention mechanism, the generated local features are input into the Long-Short Term Memory (LSTM) model, and subsequently pooled to calculate the global features of the URLs. Finally, the URLs are detected and classified by the SoftMax function using global features. The results demonstrate that compared with the existing methods, the Attention-based CNN-LSTM mechanism has higher accuracy for malicious URL detection.