• Title/Summary/Keyword: 보상인식

Search Result 440, Processing Time 0.028 seconds

Incorporation of IMM-based Feature Compensation and Uncertainty Decoding (IMM 기반 특징 보상 기법과 불확실성 디코딩의 결합)

  • Kang, Shin-Jae;Han, Chang-Woo;Kwon, Ki-Soo;Kim, Nam-Soo
    • The Journal of Korean Institute of Communications and Information Sciences
    • /
    • v.37 no.6C
    • /
    • pp.492-496
    • /
    • 2012
  • This paper presents a decoding technique for speech recognition using uncertainty information from feature compensation method to improve the speech recognition performance in the low SNR condition. Traditional feature compensation algorithms have difficulty in estimating clean feature parameters in adverse environment. Those algorithms focus on the point estimation of desired features. The point estimation of feature compensation method degrades speech recognition performance when incorrectly estimated features enter into the decoder of speech recognition. In this paper, we apply the uncertainty information from well-known feature compensation method, such as IMM, to the recognition engine. Applied technique shows better performance in the Aurora-2 DB.

DCGAN-based Compensation for Soft Errors in Face Recognition systems based on a Cross-layer Approach (얼굴인식 시스템의 소프트에러에 대한 DCGSN 기반의 크로스 레이어 보상 방법)

  • Cho, Young-Hwan;Kim, Do-Yun;Lee, Seung-Hyeon;Jeong, Gu-Min
    • The Journal of Korea Institute of Information, Electronics, and Communication Technology
    • /
    • v.14 no.5
    • /
    • pp.430-437
    • /
    • 2021
  • In this paper, we propose a robust face recognition method against soft errors with a deep convolutional generative adversarial network(DCGAN) based compensation method by a cross-layer approach. When soft-errors occur in block data of JPEG files, these blocks can be decoded inappropriately. In previous results, these blocks have been replaced using a mean face, thereby improving recognition ratio to a certain degree. This paper uses a DCGAN-based compensation approach to extend the previous results. When soft errors are detected in an embedded system layer using parity bit checkers, they are compensated in the application layer using compensated block data by a DCGAN-based compensation method. Regarding soft errors and block data loss in facial images, a DCGAN architecture is redesigned to compensate for the block data loss. Simulation results show that the proposed method effectively compensates for performance degradation due to soft errors.

Face Recognition Robust to Illumination Change (조명 변화에 강인한 얼굴 인식)

  • 류은진;박철현;구탁모;박길흠
    • Proceedings of the IEEK Conference
    • /
    • 2000.09a
    • /
    • pp.465-468
    • /
    • 2000
  • 얼굴 영상은 똑같은 표정의 같은 사람이라도 조명에 따라 매우 다른 얼굴 영상으로 나타난다. 따라서 본 논문에서는 조명 변화에 강인한 얼굴 인식 방법을 제안한다. 제안된 방법은 오프라인 훈련(off-line training)과 온라인 인식(on-line recognition)의 두 부분으로 이루어져 있다. 오프라인 훈련은 PCA(principal component analysis)를 기반으로 한다. 온라인 인식에서는 조명 변화에 대한 보상, 얼굴 특징의 추출, 그리고 인식을 위한 분류 과정의 3 단계로 구성되어 있다. 오프라인 훈련에서는 전체 훈련 얼굴 영상 데이터에 PCA를 적용하여 조명 변화가 최대한 제외된 특징 벡터 공간을 생성한다. 실제 인식 단계에서는 첫 번째로 입력 영상으로 들어온 얼굴 영상에서 조명의 영향을 보상하기 위해 준동형 필터링(homomorphic filtering) 후 밝기 정규화(normalization)를 취한다. 두 번째 단계에서는 입력 데이터의 차원을 줄이고 얼굴 특징 벡터를 구하기 위해 PCA를 수행한다. 마지막 과정으로서 입력 영상의 특징 벡터들과 오프라인에서 미리 구하여진 특징 벡터들의 유사도를 측정하여 얼굴을 인식하게 된다. 실험 결과 제안된 방법은 기존의 Eigenface 방법에 비해 우수한 성능을 나타내었다.

  • PDF

Influence of identifiable victim effect on third-party's punishment and compensation judgments (인식 가능한 피해자 효과가 제3자의 처벌 및 보상 판단에 미치는 영향)

  • Choi, InBeom;Kim, ShinWoo;Li, Hyung-Chul O.
    • Korean Journal of Forensic Psychology
    • /
    • v.11 no.2
    • /
    • pp.135-153
    • /
    • 2020
  • Identifiable victim effect refers to the tendency of greater sympathy and helping behavior to identifiable victims than to abstract, unidentifiable ones. This research tested whether this tendency also affects third-party's punishment and compensation judgments in jury context for public's legal judgments. In addition, through the Identifiable victim effect in such legal judgment, we intended to explain the effect of 'the bill named for victim', putting the victim's real name and identity at the forefront, which is aimed at strengthening the punishment of related crimes by gaining public attention and support. To do so, we conducted experiments with hypothetical traffic accident scenarios that controlled legal components while manipulating victim's identifying information. In experiment 1, each participant read a scenario of an anonymous victim (unidentifiable condition) or a nonanonymous victim that included personal information such as name and age (identifiable condition) and made judgments on the degree of punishment and compensation. The results showed no effect of identifiability on third-party's punishment and compensation judgments, but moderation effect of BJW was obtained in the identifiable condition. That is, those with higher BJW showed greater tendency of punishment and compensation for identifiable victims. In Experiment 2, we compared an anonymous victim (unidentifiable condition) against a well-conducted victim (positive condition) and ill-conducted victim (negative condition) to test the effects of victim's characteristics on punishment for offender and compensation for victims. The results showed lower compensation for an ill-conducted victim than for an anonymous one. In addition, across all conditions except for negative condition, participants made punishment and compensation judgments higher than the average judicial precedents of 10-point presented in the rating scale. This research showed that victim's characteristics other than legal components affects third-party's legal decision making. Furthermore, we interpreted third-party's tendency to impose higher punishment and compensation with effect of 'the bill named for victim' and proposed social and legal discussion for and future research.

  • PDF

On the Enhancement of the Recognition Performance for Back Propagation Neural Networks (역전파 선경회로망의 인식성능 향상에 관한 연구)

  • 홍봉화;이지영
    • Journal of the Korea Society of Computer and Information
    • /
    • v.4 no.4
    • /
    • pp.86-93
    • /
    • 1999
  • This paper proposes the multi-modular neural network and compensative input algorithm. The former is to reduce convergence speed which is one of the neural network's inveterate problems, and the latter is to improve the recognition performance of the neural network. This paper consists of two major parts and a simulation. First, it shows the structure of mu1ti-modular neural network, which is applied to the recognition of Korean, English characters and numbers. Second, it describes the compensative input algorithm and shows the steps that determine the compensative input. The proposed algorithm was tested and compared with the existing neural networks in the recognition of Korean and English characters and numbers. The convergence speed is three times or more faster than the existing neural network. In the case that compensative input was applied to neural network, the recognition rate was improved more than 10%.

  • PDF

PCA-based Variational Model Composition Method for Roust Speech Recognition with Time-Varying Background Noise (시변 잡음에 강인한 음성 인식을 위한 PCA 기반의 Variational 모델 생성 기법)

  • Kim, Wooil
    • Journal of the Korea Institute of Information and Communication Engineering
    • /
    • v.17 no.12
    • /
    • pp.2793-2799
    • /
    • 2013
  • This paper proposes an effective feature compensation method to improve speech recognition performance in time-varying background noise condition. The proposed method employs principal component analysis to improve the variational model composition method. The proposed method is employed to generate multiple environmental models for the PCGMM-based feature compensation scheme. Experimental results prove that the proposed scheme is more effective at improving speech recognition accuracy in various SNR conditions of background music, compared to the conventional front-end methods. It shows 12.14% of average relative improvement in WER compared to the previous variational model composition method.

A Study on Voice Recognition using Noise Cancel DTW for Noise Environment (잡음환경에서의 Noise Cancel DTW를 이용한 음성인식에 관한 연구)

  • Ahn, Jong-Young;Kim, Sung-Su;Kim, Su-Hoon;Koh, Si-Young;Hur, Kang-In
    • The Journal of the Institute of Internet, Broadcasting and Communication
    • /
    • v.11 no.4
    • /
    • pp.181-186
    • /
    • 2011
  • In this paper, we propose the Noise Cancel DTW that to use a kind of feature compensation. This method is not to use estimated noise but we use real life environment noise data for Voice Recognition. And we applied this contaminated data for recognition reference model that suitable for noise environment. NCDTW is combined with surround noise when generating reference patten. We improved voice recognition rate at mobile environment to use NCDTW.

Effects of Reward Programs on Brand Loyalty in Online Shopping Contexts (인터넷쇼핑 상황에서 보상프로그램이 브랜드충성도에 미치는 영향에 관한 연구)

  • Kim, Ji-Hern;Kang, Hyunmo;Munkhbazar, M.
    • Asia Marketing Journal
    • /
    • v.14 no.2
    • /
    • pp.39-63
    • /
    • 2012
  • Previous studies of reward programs have generally focused on designing the best programs for consumers and suggested that consumers' perception of the value of reward programs can vary according to the type of reward program (e.g., hedonic vs. utilitarian and direct vs. indirect) and its timing (e.g., immediate vs. delayed). These studies have typically assumed that consumers' preference for reward programs has a positive effect on brand loyalty. However, Dowling and Uncles (1997) pointed out that this preference does not necessarily foster brand loyalty. In this regard, the present study verifies this assumption by examining the effects of consumers' perception of the value of reward programs on their brand loyalty. Although reward programs are widely used by online shopping malls, most studies have examined the conditions under which consumers are most likely to value loyalty programs in the context of offline shopping. In the context of online shopping, however, consumers' preferences may have little effect on their brand loyalty because they have more opportunities for comparing diverse reward programs offered by many online shopping malls. That is, in online shopping, finding attractive reward programs may require little effort on the part of consumers, who are likely to switch to other online shopping malls. Accordingly, this study empirically examines whether consumers' perception of the value of reward programs influences their brand loyalty in the context of online shopping. Meanwhile, consumers seek utilitarian and/or hedonic value from their online shopping activity(Jones et al., 2006; Barbin et al., 1994). They visit online shopping malls to buy something necessary (utilitarian value) and/or enjoy the process of shopping itself (hedonic value). In this sense, reward programs may reinforce utilitarian as well as hedonic value, and their effect may vary according to the type of reward (utilitarian vs. hedonic). According to Chaudhuri and Holbrook (2001), consumers' perception of the value of a brand can influence their brand loyalty through brand trust and affect. Utilitarian value influences brand loyalty through brand trust, whereas hedonic value influences it through brand affect. This indicates that the effect of this perception on brand trust or affect may be moderated by the type of reward program. Specifically, this perception may have a greater effect on brand trust for utilitarian reward programs than for hedonic ones, whereas the opposite may be true for brand affect. Given the above discussion, the present study is conducted with three objectives in order to provide practical implications for online shopping malls to strategically use reward program for establishing profitable relationship with customers. First, the present study examines whether reward programs can be an effective marketing tool for increasing brand loyalty in the context of online shopping. Second, it investigates the paths through which consumers' perception of the value of reward programs influences their brand loyalty. Third, it analyzes the effects of this perception on brand trust and affect by considering the type of reward program as a moderator. This study suggests and empirically analyzes a new research model for examining how consumers' perception of the value of reward programs influences their brand loyalty in the context of online shopping. The model postulates the following 10 hypotheses about the structural relationships between five constructs: (H1) Consumers' perception of the value of reward programs has a positive effect on their program loyalty; (H2) Program loyalty has a positive effect on brand loyalty; (H3) Consumers' perception of the value of reward programs has a positive effect on their brand trust; (H4) Consumers' perception of the value of reward programs has a positive effect on their brand affect; (H5) Brand trust has a positive effect on program loyalty; (H6) Brand affect has a positive effect on program loyalty; (H7) Brand trust has a positive effect on brand loyalty; (H8) Brand affect has a positive effect on brand loyalty; (H9) Consumers' perception of the value of reward programs is more likely to influence their brand trust for utilitarian reward programs than for hedonic ones; and (H10) Consumers' perception of the value of reward programs is more likely to influence their brand affect for hedonic reward programs than for utilitarian ones. To test the hypotheses, we considered a sample of 220 undergraduate students in Korea (male:113). We randomly assigned these participants to one of two groups based on the type of reward program (utilitarian: transportation card, hedonic: movie ticket). We instructed the participants to imagine that they were offered these reward programs while visiting an online shopping mall. We then asked them to answer some questions about their perception of the value of the reward programs, program loyalty, brand loyalty, brand trust, and brand affect, in that order. We also asked some questions about their demographic backgrounds and then debriefed them. We employed the structural equation modeling (SEM) method with AMOS 18.0. The results provide support for some hypotheses (H1, H3, H4, H7, H8, and H9) while providing no support for others (H2, H5, H6, H10) (see Figure 1). Noteworthy is that the path proposed by previous studies, "value perception → program loyalty → brand loyalty," was not significant in the context of online shopping, whereas this study's proposed path, "value perception → brand trust/brand affect → brand loyalty," was significant. In addition, the results indicate that the type of reward program moderated the relationship between consumers' value perception and brand trust but not the relationship between their value perception and brand affect. These results have some important implications. First, this study is one of the first to examine how consumers' perception of the value of reward programs influences their brand loyalty in the context of online shopping. In particular, the results indicate that the proposed path, "value perception → brand trust/brand affect → brand loyalty," can better explain the effects of reward programs on brand loyalty than existing paths. Furthermore, these results suggest that online shopping malls should place greater emphasis on the type of reward program when devising reward programs. To foster brand loyalty, they should reinforce the type of shopping value that consumers emphasize by providing them with appropriate reward programs. If consumers prefer utilitarian value to hedonic value, then online shopping malls should offer utilitarian reward programs and vice versa.

  • PDF

A Study about Circumstances and Perception of Construction Accident Compensation (건설업 재해처리에 대한 환경 및 인식변화 고찰)

  • Hong, Sung-Ho;Choi, Jin-Woo
    • Journal of the Korea Institute of Building Construction
    • /
    • v.10 no.5
    • /
    • pp.69-74
    • /
    • 2010
  • It is the law that injured workers must be paid compensation through occupational health and safety insurance. But sometimes, construction companies cover the compensation for an accident by themselves in order to conceal the fact of an accident. The reason for this concealment is that they wish to get a high mark in PQ. From the point of view of a subcontractor, another reason is to avoid any negative consequences in bids. According to interviewees, some workers, knowing the vulnerability of their employers in this area, demand unreasonable levels of compensation. This study examined the circumstances, influences and participants' perception of accidents on construction sites.

Head Pose Estimation Using Error Compensated Singular Value Decomposition for 3D Face Recognition (3차원 얼굴 인식을 위한 오류 보상 특이치 분해 기반 얼굴 포즈 추정)

  • 송환종;양욱일;손광훈
    • Journal of the Institute of Electronics Engineers of Korea SP
    • /
    • v.40 no.6
    • /
    • pp.31-40
    • /
    • 2003
  • Most face recognition systems are based on 2D images and applied in many applications. However, it is difficult to recognize a face when the pose varies severely. Therefore, head pose estimation is an inevitable procedure to improve recognition rate when a face is not frontal. In this paper, we propose a novel head pose estimation algorithm for 3D face recognition. Given the 3D range image of an unknown face as an input, we automatically extract facial feature points based on the face curvature. We propose an Error Compensated Singular Value Decomposition (EC-SVD) method based on the extracted facial feature points. We obtain the initial rotation angle based on the SVD method, and perform a refinement procedure to compensate for remained errors. The proposed algorithm is performed by exploiting the extracted facial features in the normaized 3D face space. In addition, we propose a 3D nearest neighbor classifier in order to select face candidates for 3D face recognition. From simulation results, we proved the efficiency and validity of the proposed algorithm.