• Journal of information and communication convergence engineering
• /
• v.18 no.1
• /
• pp.33-38
• /
• 2020

Elasticsearch is an open source search and analytics engine that can search petabytes of data in near real time. It is designed as a distributed system horizontally scalable and highly available. It provides RESTful APIs, thereby making it programming-language agnostic. Full text search of multilingual text requires language-specific analyzers and field mappings appropriate for indexing and searching multilingual text. Additionally, a language detector can be used in conjunction with the analyzers to improve the multilingual text search. Elasticsearch provides more than 40 language analysis plugins that can process text and extract language-specific tokens and language detector plugins that can determine the language of the given text. This study investigates three different approaches to index and search Chinese, Japanese, and Korean (CJK) text (single analyzer, multi-fields, and language detector-based), and identifies the advantages of the language detector-based approach compared to the other two.

• JSTS:Journal of Semiconductor Technology and Science
• /
• v.15 no.2
• /
• pp.255-266
• /
• 2015

In this paper, a low-complexity and low-power soft output multiple input multiple output (MIMO) symbol detector is proposed for mobile devices with two transmit and two receive antennas. The proposed symbol detector can support both the spatial multiplexing mode and spatial diversity mode in single hardware and shows the optimal maximum likelihood (ML) performance. By applying a multi-stage pipeline structure and using a complex multiplier based on the polar-coordinate, the complexity of the proposed architecture is dramatically decreased. Also, by applying a clock-gating scheme to the internal modules for MIMO modes, the power consumption is also reduced. The proposed symbol detector was designed using a hardware description language (HDL) and implemented using a 65nm CMOS standard cell library. With the proposed architecture, the proposed MIMO detector takes up an area of approximately $0.31mm^2$ with 183K equivalent gates and achieves a 150Mbps throughput. Also, the power estimation results show that the proposed MIMO detector can reduce the power consumption by a maximum of 85% for the various test cases.

• Journal of KIISE:Software and Applications
• /
• v.36 no.1
• /
• pp.82-91
• /
• 2009

Sign language spotting is the task of detecting and recognizing the signs in a signed utterance. The difficulty of sign language spotting is that the occurrences of signs vary in both motion and shape. Moreover, the signs appear within a continuous gesture stream, interspersed with transitional movements between signs in a vocabulary and non-sign patterns(which include out-of-vocabulary signs, epentheses, and other movements that do not correspond to signs). In this paper, a novel method for designing a threshold model in a conditional random field(CRF) model is proposed. The proposed model performs an adaptive threshold for distinguishing between signs in the vocabulary and non-sign patterns. A hand appearance-based sign verification method, a short-sign detector, and a subsign reasoning method are included to further improve sign language spotting accuracy. Experimental results show that the proposed method can detect signs from continuous data with an 88% spotting rate and can recognize signs from isolated data with a 94% recognition rate, versus 74% and 90% respectively for CRFs without a threshold model, short-sign detector, subsign reasoning, and hand appearance-based sign verification.

• Journal of Sensor Science and Technology
• /
• v.33 no.2
• /
• pp.78-85
• /
• 2024

We proposed a method for visual place recognition that represents images using objects as visual words. Visual words represent the various objects present in urban environments. To detect various objects within the images, we implemented and used a zero-shot detector based on a large-scale image language model. This zero-shot detector enables the detection of various objects in urban environments without additional training. In the process of creating histograms using the proposed method, frequency-based weighting was applied to consider the importance of each object. Through experiments with open datasets, the potential of the proposed method was demonstrated by comparing it with another method, even in situations involving environmental or viewpoint changes.

• Annual Conference on Human and Language Technology
• /
• 2021.10a
• /
• pp.293-299
• /
• 2021

본 논문은 외국어 학습을 위한 딥러닝 기반 영어 교육 플랫폼인 PEEP-Talk (Personalized English Education Platform)을 제안한다. PEEP-Talk는 딥러닝 기반 페르소나 대화 시스템과 영어 문법 교정 피드백 기능이 내장된 교육용 플랫폼이다. 또한 기존 페르소나 대화시스템과 다르게 대화의 흐름이 벗어날 시 이를 자동으로 판단하여 대화 주제를 실시간으로 변경할 수 있는 CD (Context Detector) 모듈을 제안하며 이를 적용하여 실제 사람과 대화하는 듯한 느낌을 사용자에게 줄 수 있다. 본 논문은 PEEP-Talk의 각 모듈에 대한 정량적인 분석과 더불어 CD 모듈을 객관적으로 판단할 수 있는 새로운 성능 평가지표인 CDM (Context Detector Metric)을 기반으로 PEEP-Talk의 강건함을 검증하였다. 이와 더불어 PEEP-Talk를 카카오톡 채널을 이용하여 배포하였다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.2
• /
• pp.631-653
• /
• 2020

Sign language is a natural, visually oriented and non-verbal communication channel between people that facilitates communication through facial/bodily expressions, postures and a set of gestures. It is basically used for communication with people who are deaf or hard of hearing. In order to understand such communication quickly and accurately, the design of a successful sign language translation system is considered in this paper. The proposed system includes object detection and classification stages. Firstly, Single Shot Multi Box Detection (SSD) architecture is utilized for hand detection, then a deep learning structure based on the Inception v3 plus Support Vector Machine (SVM) that combines feature extraction and classification stages is proposed to constructively translate the detected hand gestures. A sign language fingerspelling dataset is used for the design of the proposed model. The obtained results and comparative analysis demonstrate the efficiency of using the proposed hybrid structure in sign language translation.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.47 no.3
• /
• pp.41-50
• /
• 2010

In this paper, an area-efficient symbol detector is proposed for MIMO communication systems with two transmit and two receive antennas. The proposed symbol detector can support both the spatial multiplexing mode and spatial diversity mode for MIMO transmission technique, and shows the optimal maximum likelihood (ML) performance. Also, by sharing the hardware block with multi-stage pipeline structure and using the complex multiplier based on polar-coordinate，the complexity of the proposed architecture is dramatically decreased. The proposed symbol detector was designed in hardware description language (HDL) and implemented with Xilinx Virtex-5 FPGA. With the proposed architecture, the number of logic slices for the proposed symbol detection is 52490 and the number of DSP48s (dedicated multiplier) is 52, which are reduced by 35.3% and 85.3%, respectively, compared with the conventional architecture.

• Journal of Advanced Navigation Technology
• /
• v.14 no.2
• /
• pp.220-226
• /
• 2010

In this paper, an low power symbol detector is proposed for MIMO communication system with two transmit and two receive antennas. The proposed symbol detector can support both the spatial multiplexing (SM) mode and spatial diversity (SD) mode for MIMO transmission technique, and shows the optimal maximum likelihood (ML) performance. Also, by sharing the hardware block and using the dedicated clock MIMO modes, the power of the proposed architecture is dramatically decreased. The proposed symbol detector was designed in hardware description language (HDL) and synthesized to logic gates using a $0.13-{\mu}m$ CMOS standard cell library. The power consumption was estimated by using Synopsys Power CompilerTM, which is reduced by maximum 85%, compared with the conventional architecture.

• Annual Conference on Human and Language Technology
• /
• 1995.10a
• /
• pp.112-116
• /
• 1995

본 논문에서는 한-일 대화체 기계번역 시스템을 위한 대화체 모델을 제시한다. 이 대화체 모델에서는 구문분석과 의미분석을 거치지 않고 형태소 분석만을 이용하여 대화체 모델을 구현하였다. 대화체모델은 담화문으로부터 목표를 추출하는 GOAL DETECTOR, 추출된 목표에 맞는 플랜을 제시하는 PROPOSER, 제시된 플랜의 적합성 여부를 결정하는 PROJECTOR, 플랜의 실행 후 결과를 시스템의 환경에 반영하는 EXECUTOR 및 영역에 대한 지식을 표현하는 영역지식(Domain Knowledge)으로 구성이 된다.

• International journal of advanced smart convergence
• /
• v.12 no.4
• /
• pp.171-176
• /
• 2023

This study is about a speech recognition error correction system designed to detect and correct speech recognition errors before natural language processing to increase the success rate of intent analysis in natural language processing with optimal efficiency in various service domains. An encoder is constructed to embedded the correct speech token and one or more error speech tokens corresponding to the correct speech token so that they are all located in a dense vector space for each correct token with similar vector values. One or more utterance tokens within a preset Manhattan distance based on the correct utterance token in the dense vector space for each embedded correct utterance token are detected through an error detector, and the correct answer closest to the detected error utterance token is based on the Manhattan distance. Errors are corrected by extracting the utterance token as the correct answer.