• Speech Sciences
• /
• v.10 no.3
• /
• pp.225-239
• /
• 2003

In this paper, we develop a keyword spotting system using vocabulary-independent speech recognition technique, and investigate several non-keyword modeling and post-processing methods to improve its performance. In order to model non-keyword speech segments, monophone clustering and Gaussian Mixture Model (GMM) are considered. We employ likelihood ratio scoring method for the post-processing schemes to verify the recognition results, and filler models, anti-subword models and N-best decoding results are considered as an alternative hypothesis for likelihood ratio scoring. We also examine different methods to construct anti-subword models. We evaluate the performance of our system on the automatic telephone exchange service task. The results show that GMM-based non-keyword modeling yields better performance than that using monophone clustering. According to the post-processing experiment, the method using anti-keyword model based on Kullback-Leibler distance and N-best decoding method show better performance than other methods, and we could reduce more than 50% of keyword recognition errors with keyword rejection rate of 5%.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1995.06a
• /
• pp.83-87
• /
• 1995

Keyword Spotting 이란 음성인식의 한 분야로서 입력된 음성에서 미리 정해진 특정단어 또는 복수 개의 단어들 중 어느 것이 포함되어 있는지의 여부를 찾아내고 이 단어를 식별해 내는 작업을 의미한다. 음소모델을 이용하여 Keyword Spotting 시스템을 구성할 경우 새로운 keyword의 추가 또는 변경이 필요할 때 단순히 그 발음사전에 따라 음소모델들을 연결시킴으로써 keyword 모델을 구성할 수 있으므로 단어모델에 의한 방법에 비해 장점이 있다. 본 논문에서는 triphone을 기본단위로 하는 HMM 에 의해 keyword 모델을 구성하고, non-keyword 모델 및 silence 모델을 함께 사용하는 keyword spotting 시스템을 구성하였다. 이러한 시스템에서 non-keyword 모델은 keyword와 keyword가 아닌 음성을 구분 지어주는 역할을 하므로 인식성능의 향상을 위해서는 적절한 non-keyword 모델의 선택이 필요하다. 본 논문에서는 10개의 state를 갖는 단일모델, 조음방법에 의해 음소들을 clustering 한 모델, 그리고 통계적 방법에 의해 음소들을 clustering 한 모델들을 각각 non-keyword 모델로 사용하여 그 성능을 비교하였다. 6개의 keyword를 대상으로 한 화자독립 keyword spotting 실험결과, 통계적 방법에 의해 음소들을 6 또는 7개의 그룹으로 clustering 한 방법이 가장 우수한 인식성능을 나타냈다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.10 no.3
• /
• pp.421-429
• /
• 2006

Recently, DHT-based Structured P2P System have a attention to scalability and providing efficient lookup by routing. Retrieving content of DHT-based P2P require knowledge of the exact identifier of sharing file. But user may wish to search for sharing file using descriptive keyword or content. To resolve the problem, this paper propose Extended-CAN mechanism to support keyword search. We defined content-based keyword and identifier of sharing file, and PLS extended to KID and CKD for keyword, common keyword processing. As a result, Extended-CAN mechanism provide efficient keyword search for DHT-based Structured P2P System.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.5
• /
• pp.2718-2731
• /
• 2019

Conjunctive keyword search encryption is an important technique for protecting sensitive data that is outsourced to cloud servers. However, the process of searching outsourced data may facilitate the leakage of sensitive data. Thus, an efficient data search approach with high security is critical. To solve this problem, an efficient conjunctive keyword search scheme based on ciphertext-policy attribute-based encryption is proposed for cloud storage environment. This paper proposes an efficient mechanism for removing the secure channel and resisting off-line keyword-guessing attacks. The storage overhead and the computational complexity are regardless of the number of keywords. This scheme is proved adaptively secure based on the decisional bilinear Diffie-Hellman assumption in the standard model. Finally, the results of theoretical analysis and experimental simulation show that the proposed scheme has advantages in security, storage overhead and efficiency, and it is more suitable for practical applications.

• Journal of the Korea Convergence Society
• /
• v.9 no.7
• /
• pp.17-24
• /
• 2018

One of the major issues in big data processing is extracting keywords from internet and using them to process the necessary information. Most of the proposed keyword extraction algorithms extract keywords using search function of a large portal site. In addition, these methods extract keywords based on already posted or created documents or fixed contents. In this paper, we propose a KAES(Keyword Advertisement Extraction System) system that helps the potential shopping keyword marketing to extract issue keywords and related keywords based on dynamic instant messages such as various issues, interests, comments posted on SNS. The KAES system makes a list of specific accounts to extract keywords and related keywords that have most frequency in the SNS.

• Journal of information and communication convergence engineering
• /
• v.16 no.1
• /
• pp.48-51
• /
• 2018

The traditional documents analysis was centered on words based system was implemented using a morpheme analyzer. These traditional systems can classify used words in the document but, cannot help to user's document understanding or analysis. In this problem solved, System needs extract for most valuable paragraphs what can help to user understanding documents. In this paper, we propose system extracts paragraphs of normalized XML document. User insert to system what filename when wants for analyze XML document. Then, system is search for keyword of the document. And system shows results searched keyword. When user choice and inserts keyword for user wants then, extracting for paragraph including keyword. After extracting paragraph, system operating maintenance paragraph sequence and check duplication. If exist duplication then, system deletes paragraph of duplication. And system informs result to user what counting each keyword frequency and weight to user, sorted paragraphs.

• Proceedings of the KSPS conference
• /
• 2005.11a
• /
• pp.155-158
• /
• 2005

Keyword spotting is effective to find keyword from the continuously pronounced speech. However, non-keyword may be accepted as keyword when the environmental noise occurs or speaker changes. To overcome this performance degradation, utterance rejection techniques using confidence measure on the recognition result have been developed. In this paper, we apply DTW to the HMM based broadcasting news keyword spotting system for rejecting non-keyword. Experimental result shows that false acceptance rate is decreased to 50%.

• Speech Sciences
• /
• v.3
• /
• pp.119-131
• /
• 1998

We have developed a keyword spotting system for automatic speech editing. This system recognizes the only keyword 'MBC news' and then sends the time information to the host system. We adopted a vocabulary dependent model based on continuous hidden Markov model, and the Viterbi search was used for recognizing the keyword. In recognizing the keyword, the system uses a parallel network where HMM models are connected independently and back-tracking information for reducing false alarms and missing. We especially focused on implementing a stable and practical real-time system.

• Journal of the Korea Society of Computer and Information
• /
• v.22 no.2
• /
• pp.97-104
• /
• 2017

In semantic information retrieval, we first need to build domain ontology and second, we need to convert the users' search keywords into a standard query such as SPARQL. In this paper, we propose a method that can automatically convert the users' search keywords into the SPARQL queries. Furthermore, our method can ensure effective performance in a specific domain such as law. Our method constructs the keyword history ontology by associating each keyword with a series of information when there are multiple keywords. The constructed ontology will convert keyword history ontology into SPARQL query. The automatic transformation method of SPARQL query proposed in the paper is converted into the query statement that is deemed the most appropriate by the user's intended keywords. Our study is based on the existing legal ontology constructions that supplement and reconstruct schema and use it as experiment. In addition, design and implementation of a semantic search tool based on legal domain and conduct experiments. Based on the method proposed in this paper, the semantic information retrieval based on the keyword is made possible in a legal domain. And, such a method can be applied to the other domains.

• Journal of Software Engineering Society
• /
• v.29 no.2
• /
• pp.13-20
• /
• 2020

The exploratory test was introduced as a desirable test method due to its fast development cycle, but it is not actively adopted because documentation and analysis of the test range are required for application. On the other hand, keyword-based testing has been introduced as a way to save resources and facilitate maintenance, but it is difficult to plan tests in advance due to the large number of variables such as data, settings, interactions, sequence and timing. However, in keyword-based testing, you can create a test case based on keywords by presenting clear criteria and methods for creating keywords and applying the exploration testing process. In this paper, we propose a model that automates exploratory tests based on keywords. To verify the effectiveness, we compared the general keyword-based test(KBT) and keyword-based exploratory test(KBET), and compared with the exploratory normal test case(ETC) and keyword-based exploratory test(KBET).