• Title/Summary/Keyword: state recognition

Search Result 1,016, Processing Time 0.025 seconds

Fast offline transformer-based end-to-end automatic speech recognition for real-world applications

  • Oh, Yoo Rhee;Park, Kiyoung;Park, Jeon Gue
    • ETRI Journal
    • /
    • v.44 no.3
    • /
    • pp.476-490
    • /
    • 2022
  • With the recent advances in technology, automatic speech recognition (ASR) has been widely used in real-world applications. The efficiency of converting large amounts of speech into text accurately with limited resources has become more vital than ever. In this study, we propose a method to rapidly recognize a large speech database via a transformer-based end-to-end model. Transformers have improved the state-of-the-art performance in many fields. However, they are not easy to use for long sequences. In this study, various techniques to accelerate the recognition of real-world speeches are proposed and tested, including decoding via multiple-utterance-batched beam search, detecting end of speech based on a connectionist temporal classification (CTC), restricting the CTC-prefix score, and splitting long speeches into short segments. Experiments are conducted with the Librispeech dataset and the real-world Korean ASR tasks to verify the proposed methods. From the experiments, the proposed system can convert 8 h of speeches spoken at real-world meetings into text in less than 3 min with a 10.73% character error rate, which is 27.1% relatively lower than that of conventional systems.

Trends and Future Directions in Facial Expression Recognition Technology: A Text Mining Analysis Approach (얼굴 표정 인식 기술의 동향과 향후 방향: 텍스트 마이닝 분석을 중심으로)

  • Insu Jeon;Byeongcheon Lee;Subeen Leem;Jihoon Moon
    • Annual Conference of KIPS
    • /
    • 2023.05a
    • /
    • pp.748-750
    • /
    • 2023
  • Facial expression recognition technology's rapid growth and development have garnered significant attention in recent years. This technology holds immense potential for various applications, making it crucial to stay up-to-date with the latest trends and advancements. Simultaneously, it is essential to identify and address the challenges that impede the technology's progress. Motivated by these factors, this study aims to understand the latest trends, future directions, and challenges in facial expression recognition technology by utilizing text mining to analyze papers published between 2020 and 2023. Our research focuses on discerning which aspects of these papers provide valuable insights into the field's recent developments and issues. By doing so, we aim to present the information in an accessible and engaging manner for readers, enabling them to understand the current state and future potential of facial expression recognition technology. Ultimately, our study seeks to contribute to the ongoing dialogue and facilitate further advancements in this rapidly evolving field.

Improving the Recognition of Known and Unknown Plant Disease Classes Using Deep Learning

  • Yao Meng;Jaehwan Lee;Alvaro Fuentes;Mun Haeng Lee;Taehyun Kim;Sook Yoon;Dong Sun Park
    • Smart Media Journal
    • /
    • v.13 no.8
    • /
    • pp.16-25
    • /
    • 2024
  • Recently, there has been a growing emphasis on identifying both known and unknown diseases in plant disease recognition. In this task, a model trained only on images of known classes is required to classify an input image into either one of the known classes or into an unknown class. Consequently, the capability to recognize unknown diseases is critical for model deployment. To enhance this capability, we are considering three factors. Firstly, we propose a new logits-based scoring function for unknown scores. Secondly, initial experiments indicate that a compact feature space is crucial for the effectiveness of logits-based methods, leading us to employ the AM-Softmax loss instead of Cross-entropy loss during training. Thirdly, drawing inspiration from the efficacy of transfer learning, we utilize a large plant-relevant dataset, PlantCLEF2022, for pre-training a model. The experimental results suggest that our method outperforms current algorithms. Specifically, our method achieved a performance of 97.90 CSA, 91.77 AUROC, and 90.63 OSCR with the ResNet50 model and a performance of 98.28 CSA, 92.05 AUROC, and 91.12 OSCR with the ConvNext base model. We believe that our study will contribute to the community.

Multi-Cultural Space and Glocal Ethics : From Cultural Space of Transnational Capitalism to Space of Recognition Struggle (다문화공간과 지구-지방적 윤리 : 초국적 자본주의의 문화공간에서 인정투쟁의 공간으로)

  • Choi, Byung-Doo
    • Journal of the Korean association of regional geographers
    • /
    • v.15 no.5
    • /
    • pp.635-654
    • /
    • 2009
  • Recently, concepts of multicultural society and/or multiculturalism have been not only widely discussed across several disciplines, but also actively promoted in government's policy, as the in-flow of foreign immigrants has increased rapidly. This paper suggests the term 'multicultural space' instead of multicultural society in a sense that both international migration of immigrants and their accommodation to a certain locality presuppose a spatial dimension. This paper also points out that the term multiculturalsim should be used very carefully, because this term includes a normative character implied in a sense of recognition of ethnic and cultural diversity and difference on the one hand, and an ideological one reflected on strategic policies of capital and the state on the other. On the basis of recognition of these problems, this paper tries to reformulate spatially the concept of muticultural society which has been supposed to be constructed due to rapidly increasing foreign immigrants, emphasizing some usefulness of multi-scalar approach. It then analyzes economic and political contexts of transnational migration, providing a criticism of multiculturalism as an ideological logic of capital and the state in transnational captialism. Finally it put a stress upon importance of struggle for spaces of recognition as a new glocal ethics in the age of post-globalization.

  • PDF

A Study on the Application of the New York Convention in the Recognition and Enforcement of ISDS Arbitral Awards (투자협정중재에 의한 중재판정의 승인·집행에 대한 뉴욕협약 적용에 관한 고찰)

  • Kang, Soo Mi
    • Journal of Arbitration Studies
    • /
    • v.29 no.1
    • /
    • pp.31-52
    • /
    • 2019
  • As international transactions have grown more numerous, situations of disputes related to the transactions are getting more complicated and more diverse. Cost-effective remedies to settle the disputes through traditional methods such as adjudications of a court will be insufficient. There fore, nations are attempting to more efficiently solve investor-state disputes through arbitration under organizations such as the ICSID Convention, the ICSID Additionary Facility Rules, and the UNCITRAL Arbitration Rules by including the provisions on investor-state dispute settlement at the conclusion of an investment agreement. In case of an arbitration under the ICSID Convention, ICSID directly exercises the supervisorial function on arbitral proceedings, and there is no room for the intervention of national courts. In time of the arbitration where the ICSID Convention does not apply, however, the courts have to facilitate the arbitral proceedings. When the recognition and enforcement of an arbitral award under the ICSID Convention are guaranteed by the Convention, it should be considered that the New York Convention does not apply to them under the Convention Article 7 (1) fore-end. In exceptional cases in which an arbitral award under the ICSID Convention cannot be recognized or enforced by the Convention, the New York Convention applies to the recognition and enforcement because the award is not a domestic award of the country in which the recognition or enforcement is sought. It is up to an interpretation of the New York Convention whether the New York Convention applies to ISDS arbitral awards not based on the ICSID Convention or not. Although an act of the host country is about sovereign activities, a host country and the country an investor is in concurring to the investment agreement with the ISDS provisions is considered a surrender of sovereignty immunity, and it will not suffice to exclude the investment disputes from the scope of application of the New York Convention. If the party to the investment agreement has declared commercial reservation at its accession into the New York Convention, it should be viewed that the Convention applies to the recognition and enforcement of the ISDS awards to settle the disputes over an investitive act, inasmuch as the act will be considered as a commercial transaction. When the recognition and enforcement of an arbitral award on investment disputes about a nation's sovereign act have been sought in Korea and Korea has been designated the place of the investment agreement arbitration as a third country, it should be reviewed whether the disputes receive arbitrability under the Korean Arbitration Act or not.

Development of a Korean Speech Recognition Platform (ECHOS) (한국어 음성인식 플랫폼 (ECHOS) 개발)

  • Kwon Oh-Wook;Kwon Sukbong;Jang Gyucheol;Yun Sungrack;Kim Yong-Rae;Jang Kwang-Dong;Kim Hoi-Rin;Yoo Changdong;Kim Bong-Wan;Lee Yong-Ju
    • The Journal of the Acoustical Society of Korea
    • /
    • v.24 no.8
    • /
    • pp.498-504
    • /
    • 2005
  • We introduce a Korean speech recognition platform (ECHOS) developed for education and research Purposes. ECHOS lowers the entry barrier to speech recognition research and can be used as a reference engine by providing elementary speech recognition modules. It has an easy simple object-oriented architecture, implemented in the C++ language with the standard template library. The input of the ECHOS is digital speech data sampled at 8 or 16 kHz. Its output is the 1-best recognition result. N-best recognition results, and a word graph. The recognition engine is composed of MFCC/PLP feature extraction, HMM-based acoustic modeling, n-gram language modeling, finite state network (FSN)- and lexical tree-based search algorithms. It can handle various tasks from isolated word recognition to large vocabulary continuous speech recognition. We compare the performance of ECHOS and hidden Markov model toolkit (HTK) for validation. In an FSN-based task. ECHOS shows similar word accuracy while the recognition time is doubled because of object-oriented implementation. For a 8000-word continuous speech recognition task, using the lexical tree search algorithm different from the algorithm used in HTK, it increases the word error rate by $40\%$ relatively but reduces the recognition time to half.

Bicycle Riding-State Recognition Using 3-Axis Accelerometer (3축 가속도센서를 이용한 자전거의 주행 상황 인식 기술 개발)

  • Choi, Jung-Hwan;Yang, Yoon-Seok;Ru, Mun-Ho
    • Journal of the Institute of Electronics Engineers of Korea SC
    • /
    • v.48 no.6
    • /
    • pp.63-70
    • /
    • 2011
  • A bicycle is different from vehicles in the structure that a rider is fully exposed to the surrounding environment. Therefore, it needs to make use of prior information about local weather, air quality, trail road condition. Moreover, since it depends on human power for moving, it should acquire route property such as hill slope, winding, and road surface to improve its efficiency in everyday use. Recent mobile applications which are to be used during bicycle riding let us aware of the necessity of development of intelligent bicycles. This study aims to develop a riding state (up-hill, down-hill, accelerating, braking) recognition algorithm using a low-power wrist watch type embedded system which has 3-axis accelerometer and wireless communication capability. The developed algorithm was applied to 19 experimental riding data and showed more than 95% of correct recognition over 83.3% of the total dataset. The altitude and temperature sensor also in the embedded system mounted on the bicycle is being used to improve the accuracy of the algorithm. The developed riding state recognition algorithm is expected to be a platform technology for intelligent bicycle interface system.

A Study on a Model Parameter Compensation Method for Noise-Robust Speech Recognition (잡음환경에서의 음성인식을 위한 모델 파라미터 변환 방식에 관한 연구)

  • Chang, Yuk-Hyeun;Chung, Yong-Joo;Park, Sung-Hyun;Un, Chong-Kwan
    • The Journal of the Acoustical Society of Korea
    • /
    • v.16 no.5
    • /
    • pp.112-121
    • /
    • 1997
  • In this paper, we study a model parameter compensation method for noise-robust speech recognition. We study model parameter compensation on a sentence by sentence and no other informations are used. Parallel model combination(PMC), well known as a model parameter compensation algorithm, is implemented and used for a reference of performance comparision. We also propose a modified PMC method which tunes model parameter with an association factor that controls average variability of gaussian mixtures and variability of single gaussian mixture per state for more robust modeling. We obtain a re-estimation solution of environmental variables based on the expectation-maximization(EM) algorithm in the cepstral domain. To evaluate the performance of the model compensation methods, we perform experiments on speaker-independent isolated word recognition. Noise sources used are white gaussian and driving car noise. To get corrupted speech we added noise to clean speech at various signal-to-noise ratio(SNR). We use noise mean and variance modeled by 3 frame noise data. Experimental result of the VTS approach is superior to other methods. The scheme of the zero order VTS approach is similar to the modified PMC method in adapting mean vector only. But, the recognition rate of the Zero order VTS approach is higher than PMC and modified PMC method based on log-normal approximation.

  • PDF

AN ASSET MANAGEMENT ASSESSMENT MODEL FOR STATE DOTs

  • Steven Cooksey;Hyung Seok David Jeong;Myung-Jin Chae
    • International conference on construction engineering and project management
    • /
    • 2009.05a
    • /
    • pp.380-387
    • /
    • 2009
  • In the past, many state Departments of Transportation (DOTs) in the U.S. managed their highway assets on a "worst first" basis and planned their highway projects in a tactical rather than strategic fashion. Due to increasingly tight highway budgets and recognition of long term benefits of asset management systems, the Federal Highway Administration (FHWA) has strongly pushed and encouraged state DOTs to implement asset management for managing their highway assets and highway projects. Currently, many DOTs have actively implemented and are in the process of applying this asset management concept for their highway infrastructure. However, different DOTs are developing different asset management systems because of their different organizational structures, data management structures, relationship with the legislature, and investment priorities. This study first identifies asset management indicators which are essential to successfully implementing asset management systems for State highway assets. The research team conducted a survey of asset management experts and reviewed the practices and policies of leading DOTs in asset management. Based on these indicators, this study develops an Asset Management Assessment Model (AM2) for different asset management systems. This model can be used by different DOTs to evaluate their current asset management systems and identify their strong areas and also their weak areas to improve in order to fully benefit from the advanced concept of asset management.

  • PDF

Seeing the State-nature Relation in South Korea from the Perspective of Political Ecology (한국의 국가와 자연의 관계에 대한 정치생태학적 연구를 위한 시론)

  • Hwang, Jin-Tae;Park, Bae-Gyoon
    • Journal of the Korean Geographical Society
    • /
    • v.48 no.3
    • /
    • pp.348-365
    • /
    • 2013
  • This paper aims to examine the complexities of the state-nature relations in Korea by emphasizing the complex processes of interactions between the state and nature. In doing so, it relies on the literature of "political ecology of state-nature" which problematizes the conventional modernist views on nature assuming the dualistic separation between the state and nature. First, we critically review the existing Korean literature on the state-nature relation (e.g., the ecologism, the metabolic rift theory, the social construction of the nature, the green state thesis, etc.) and argue that these studies significantly lack the recognition of the interactions between the state and nature. Second, we discuss the possibilities of seeing the state-nature relations from the perspective of political ecology as an alternative approach to the state-nature relation. Last, we conclude that the political ecology approach to the state-nature can deepen our understandings of the Korean capitalist development.

  • PDF