• 한국콘텐츠학회논문지
• /
• 제5권5호
• /
• pp.59-64
• /
• 2005

홍채 인식은 판별인자가 지문에 비해 매우 높은 정확도를 갖는다. 홍채의 주름을 주파수로 바꾸는 과정을 통해 짧은 시간 내에 인식 가능하며 살아있는 사람의 흥채는 미세한 떨림이 있기 때문에 도용이 거의 불가능하다. 하지만 홍채의 영상은 시간이 지나 인체의 변화에 따라 홍채가 변화될 경우 기존의 흥채를 이용한 신원 확인시스템은 오인식할 수 있다는 문제점이 발생할 수 있다. 본 논문에서는 신원확인 시스템에서 생체 인식을 위한 RIAA(Renewable Iris Authentication Algorithm) 알고리즘을 제안하고 구현하였다. 이 알고리즘은 신원 확인을 위한 홍채 인식 방법에 관한 것으로서, 홍채를 일정층상의 측면에서 단층 촬영할 때 나타나는 등고선 즉, 돌출 혹은 침강 면의 경계선을 근거로 홍채코드를 생성하여 원본과 비교하게 함으로서 개인의 신원을 확인하도록 하는 홍채 인식방법에 관한 것이다.

• 한국멀티미디어학회논문지
• /
• 제22권6호
• /
• pp.655-664
• /
• 2019

In this paper, we propose a novel way of combining multiple deep convolutional neural network (DCNN) architectures which work well for accurate video face identification by adopting a serial combination of 3D and 2D DCNNs. The proposed method first divides an input video sequence (to be recognized) into a number of sub-video sequences. The resulting sub-video sequences are used as input to the 3D DCNN so as to obtain the class-confidence scores for a given input video sequence by considering both temporal and spatial face feature characteristics of input video sequence. The class-confidence scores obtained from corresponding sub-video sequences is combined by forming our proposed class-confidence matrix. The resulting class-confidence matrix is then used as an input for learning 2D DCNN learning which is serially linked to 3D DCNN. Finally, fine-tuned, serially combined DCNN framework is applied for recognizing the identity present in a given test video sequence. To verify the effectiveness of our proposed method, extensive and comparative experiments have been conducted to evaluate our method on COX face databases with their standard face identification protocols. Experimental results showed that our method can achieve better or comparable identification rate compared to other state-of-the-art video FR methods.

• 한국정보과학회논문지:소프트웨어및응용
• /
• 제27권5호
• /
• pp.583-592
• /
• 2000

본 논문에서는 독립성분분석을 이용한 음성의 특징 벡터 변환방법을 제안한다. 제안한 방법은 여러 환경에서 수집된 음성신호의 켑스트럼 벡터를 다수의 특징 함수들의 선형결합으로 가정하고, 독립성분분석을 이용하여 분리된 켑스트럼 벡터를 학습과 인식에 사용한다. 변환된 벡터 영역에서는 반복적으로 나타나는 화자의 특징 정보는 강조되고 임의로 나타나는 채널 왜곡은 억제되는 효과를 볼 수 있다. 제안된 방법의 유효성을 검증하기 위해 실제 전화음성으로 문장독립형 화자식별 실험을 수행하였으며, 결과를 통해 독립성분분석을 이용한 특징벡터의 변환이 채널 환경 변화에 대해 보다 강인함을 보였다.

• 말소리와 음성과학
• /
• 제2권3호
• /
• pp.91-99
• /
• 2010

This paper reports on an identification test where KFL learners identified the Korean three-way laryngeal contrast in the phrase-initial position, when the phrase-initial tone was systematically manipulated. It turns out that heritage learners have some sensitivity to phrase-initial tone and show a plain-aspirated alternation in their identification according to the phrase-initial tone, as native speakers do, whereas non-heritage students do not show such tone sensitivity. However, after a weekly prosody training, second-year non-heritage students have shown a significant improvement in their performance. This paper clearly shows that the phrase-initial tone plays a critical role in distinguishing laryngeal features of Korean obstruents, and also suggests that prosody including the tone-segment correlation should be incorporated in the KFL curriculum.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1999년도 추계학술대회 논문집 학회본부 B
• /
• pp.821-823
• /
• 1999

This paper describes a study on a person identification system which can improve currently available biometric systems. In the procedure of PIN(Personal Identification Number) input, holding time, interkey time between key presses are measured and normalized. Person identification is performed by matching using Euclidean distance of these punching dynamics. The experimental results show the possibility of improvement of the overall system performance when keypad dynamics feature is applied to the biometric systems which take PIN input using keypads.

• 말소리와 음성과학
• /
• 제4권2호
• /
• pp.79-86
• /
• 2012

In this paper, we propose a novel approach to improve histogram equalization for speaker identification. Our method collects all speech features of UBM training data to make a reference distribution. The ranks of the feature vectors are calculated in the sorted list of the collection of the UBM training data and the test data. We use the ranks to perform order-based histogram equalization. The proposed method improves the accuracy of the speaker recognition system with short utterances. We use four kinds of speech databases to evaluate the proposed speaker recognition system and compare the system with cepstral mean normalization (CMN), mean and variance normalization (MVN), and histogram equalization (HEQ). Our system reduced the relative error rate by 33.3% from the baseline system.

• ETRI Journal
• /
• 제40권5호
• /
• pp.643-652
• /
• 2018

Dysarthria is a degenerative disorder of the central nervous system that affects the control of articulation and pitch; therefore, it affects the uniqueness of sound produced by the speaker. Hence, dysarthric speaker recognition is a challenging task. In this paper, a feature-extraction method based on deep belief networks is presented for the task of identifying a speaker suffering from dysarthria. The effectiveness of the proposed method is demonstrated and compared with well-known Mel-frequency cepstral coefficient features. For classification purposes, the use of a multi-layer perceptron neural network is proposed with two structures. Our evaluations using the universal access speech database produced promising results and outperformed other baseline methods. In addition, speaker identification under both text-dependent and text-independent conditions are explored. The highest accuracy achieved using the proposed system is 97.3%.

• 제어로봇시스템학회논문지
• /
• 제18권9호
• /
• pp.801-805
• /
• 2012

This paper proposes dynamic thresholding scheme for fingerprint identification. As a user authentication method by fingerprint recognition technology, verification method based on 1:1 matching was mainly used in the past, but identification method based on 1:N matching is generally used recently. The control of the value of FAR is very important in the application areas such as access control and time attendance systems. This paper proposes dynamic thresholding scheme which could properly control the value of FAR according to the field of applications and size of the fingerprints database.

• 전기학회논문지
• /
• 제61권4호
• /
• pp.542-548
• /
• 2012

This study presents a new method of system modeling by using the Discrete Fourier Transform for the position control system of DC Motor. And the proposed method is similar to the method of System Identification by analysis of correlation of the measured input-output data. The measured output signals are transformed to the frequency domain using DFT. The Fourier Spectrum of the transformed signals is used for knowing to the feature of having an important effect on the system. And transfer function of the second order system is estimated by the dominant parameter which is computed in the magnitude and the phase of Fourier spectrum of the transformed signals. In addition, the output signal includes the unique feature of system. So, although the basic parameter of the system is unknown for us, the proposed method has an advantage to system modeling. And the controller is easily designed by the estimated transfer function. Thus, in this paper, the proposed method is applied to the system modeling for the position control system of DC Motor and the PD-controller is designed by the estimated model. And the efficiency and the reliability of the proposed method are verified by the experimental result.

• 한국콘텐츠학회논문지
• /
• 제12권3호
• /
• pp.67-76
• /
• 2012

본 논문은 천 커버링 연산의 개념과 알고리즘을 제안한다. 천 커버링은 물리 법칙에 기반을 둔 연산으로 사물을 덮는 천의 모양을 계산학적으로 흉내낸다. 천 커버링의 목적은 사물을 천으로 덮어 표면의 상세함을 감추고 개략적인 외형이 드러나도록 하는 것이다. 이 연산은 천의 뻣뻣한 정도를 제어하는 하나의 크기 인자를 가지며, 이를 통해 외부로 드러나는 사물에 대한 정보의 상세함을 조절한다. 제안하는 연산의 가능성을 보이기 위해 문서 영상에 사용된 언어를 식별하는 문제에 천 커버링을 적용하였다. 실험 결과 가우시안을 이용한 특징 추출 방법보다 천 커버링을 이용한 특징 추출 방법이 더 우수한 식별 성능을 보였다. 토론에서 제안하는 연산이 우수한 이유를 제시한다.