• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2013.10a
• /
• pp.447-450
• /
• 2013

This paper proposes the way to prevent drowsy driving by recognizing drivers eyes and yawn using a front camera. The method uses the Viola-Jones algorithm to detect eyes area and mouth area from detection face region. In the eyes area, it uses the Hough transform to recognize eye circle in order to distinguish drowsy driving. In the mouth area, it determines whether for the driver to yawn through a sub-window testing by applying a HSV-filter and detecting skin color of the tongue. The test result shows that the recognition rate of yawn reaches up to 90%. It is expected that the method introduced in this paper might contribute to reduce the number of drowsy driving accidents.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2006.05a
• /
• pp.299-302
• /
• 2006

본 논문에서는 다중필터와 복합형 신경망으로 구성된 얼굴 검출 시스템과 WFMM 신경망을 이용한 피부색 검출기법을 소개한다. 전처리 단계에 해당하는 다중필터는 대상 영역의 수를 감소 시켜 시스템의 속도를 개선한다. 다중필터에 속한 색상필터는 총 11 가지의 색상 공간에서 피부색의 특징 값을 추출하여 학습 데이터로 사용하며, 이 학습 데이터에 의해 생성된 하이퍼 박스를 통해 피부색을 분류한다. 또한 WFMM 신경망의 연관도 요소 특성을 이용하여 각 색상 공간의 상대적 중요도를 분석하여 피부색 검출에 유용한 색상 공간을 분석하고 추출 한다. 얼굴패턴 검출을 위한 복합형 신경망은 첫 단계에서 가보 변환을 사용하는 CNN 을 통해 특징 지도를 생성하고, WFMM 신경망으로 최종 얼굴패턴을 검증한다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2006.10b
• /
• pp.489-493
• /
• 2006

이 논문에서는 커널 에지 방식의 얼굴의 특징점을 추출하는 방법과 Adaboost를 이용한 얼굴의 특징점을 추출하는 방법에 대해서 비교 한다. 커널 에지를 이용한 방법은 10개의 커널을 이용하여 추출된 에지를 이용하여 얼굴의 특징점을 추출해 낸다. 커널의 개수를 줄여 사용한다면 실시간에 가능하고, 정확성을 높이기 위해서는 이미지의 전처리 단계에서 자극적인 효과를 준다면 정확성 또한 높아 질 것이다. 반면에 Adaboost를 이용한 방법은 각각의 특징점들을 오프라인 상에서 학습을 하고 온라인상에서 실시간으로 특징점을 추출하는 방법을 사용하였다. 각 각의 학습과정에 있어서 positive, negative 이미지를 더 많이 사용한다면 정확성이 더 높아질 것이다. 한 가지 주목할 만 한 점은 입과 같은 특징점을 추출하기 어려운 영역에서도 높은 정확성을 보였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2006.11a
• /
• pp.269-272
• /
• 2006

얼굴 인식의 전처리 단계로써 얼굴의 특징 영역인 눈, 코, 입을 추출하는 방법들이 최근 다양하게 연구되고 있다. 얼굴 영상의 특징 영역을 추출 하는 방법에는 일반적으로 특징 점을 이용한 방법과 에지 정보를 이용한 방법이 있다. 특징 점을 이용한 방법은 높은 정확도를 보이는 반면 느린 수행시간을 보이는 문제점이 있으며, 에지 정보를 이용한 방법은 빠른 수행시간을 보이지만 정확도가 떨어지는 문제점이 있다. 본 논문에서는 정확도와 수행시간을 동시에 향상시킬 수 있는 방법을 제안한다. 빠른 수행 시간을 위해 에지 정보와 에지의 방향성 정보를 이용하여 대략적으로 영역을 추출하여, 잡음에 의해 발생된 에지나 빛에 의해 추출되지 못한 에지에서 생긴 눈 추출의 오류는 추출된 영역의 가로, 세로 비율과 각 영역의 공간 정보를 이용하여 해결한다. 실험 결과에서 85%의 정확도와 평균 0.3초의 수행시간을 보였으며, 에지 정보를 이용한 방법의 문제점인 정확도와 특징 점을 이용한 방법의 문제점인 수행시간을 동시에 향상시킨 결과를 보였다.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.41 no.3
• /
• pp.27-36
• /
• 2004

Detection and recognition of human faces in images can be considered as an important aspect for applications that involve interaction between human and computer. In this paper, we propose a face recognition method using eigenfaces and fuzzy neural networks. The Principal Components Analysis (PCA) is one of the most successful technique that have been used to recognize faces in images. In this technique the eigenvectors (eigenfaces) and eigenvalues of an image is extracted from a covariance matrix which is constructed form image database. Face recognition is Performed by projecting an unknown image into the subspace spanned by the eigenfaces and by comparing its position in the face space with the positions of known indivisuals. Based on this technique, we propose a new algorithm for face recognition consisting of 5 steps including preprocessing, eigenfaces generation, design of fuzzy membership function, training of neural network, and recognition. First, each face image in the face database is preprocessed and eigenfaces are created. Fuzzy membership degrees are assigned to 135 eigenface weights, and these membership degrees are then inputted to a neural network to be trained. After training, the output value of the neural network is intupreted as the degree of face closeness to each face in the training database.

• The KIPS Transactions:PartB
• /
• v.13B no.1 s.104
• /
• pp.7-14
• /
• 2006

This paper presents an efficient face recognition method using both first moment of image and basis images. First moment which is a method for finding centroid of image, is applied to exclude the needless backgrounds in the face recognitions by shifting to the centroid of face image. Basis images which are the face features, are respectively extracted by principal component analysis(PCA) and fixed-point independent component analysis(FP-ICA). This is to improve the recognition performance by excluding the redundancy considering to second- and higher-order statistics of face image. The proposed methods has been applied to the problem for recognizing the 48 face images(12 persons*4 scenes) of 64*64 pixels. The 3 distances such as city-block, Euclidean, negative angle are used as measures when match the probe images to the nearest gallery images. The experimental results show that the proposed methods has a superior recognition performances(speed, rate) than conventional PCA and FP-ICA without preprocessing, the proposed FP-ICA has also better performance than the proposed PCA. The city-block has been relatively achieved more an accurate similarity than Euclidean or negative angle.

• The KIPS Transactions:PartB
• /
• v.10B no.6
• /
• pp.711-718
• /
• 2003

In this paper, we propose a robust face tracking algorithm. It is based on Condensation algorithm [7] and uses skin color and facial shape as the observation measure. It is hard to integrate color weight and shape weight. So we propose the method that has two separate trackers which uses skin color and facial shape as the observation measure respectively. One tracker tracks skin colored region and the other tracks facial shape. We used importance sampling technique to limit sampling region of two trackers. For skin-colored region tracker, we propose an adaptive color model to avoid the effect of illumination change. The proposed face tracker performs robustly in clutter background and in the illumination changes.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2017.11a
• /
• pp.781-784
• /
• 2017

Face recognition systems are characterized by low invasiveness of acquisition, and increasingly better reliability. Such systems may not be applied effectively, when the images are in low resolution (LR) as in the case that photos are taken from long distances, typically public surveillance. In theory, the high resolution (HR) image reconstructed from an LR face image, applying a super resolution (SR) method, can be used for face recognition. However, existing face SR algorithms may not give satisfactory results in face recognition. This article investigates the very low resolution face recognition problem and introduces a partial differential equation (PDE)-based SR method for a face recognition system of convolutional neural network (CNN).

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.6C
• /
• pp.570-576
• /
• 2005

Recognition of posed face is one of the most challenging problems in the field of face recognition. In this paper, as a preprocessing step for recognizing such faces, a method to transform non-frontal face images into frontal face images is proposed. The linear relationship between eigenfaces is utilized to obtain a pose transform matrix. The proposed method is verified with a well-known face recognition algorithm based on PCA/LDA. Compared to the conventional algorithm applied to the original posed face images, our experimental results indicated that the proposed method contributes to improve the recognition rate of such faces by $20\%$.

• Journal of Korea Multimedia Society
• /
• v.10 no.2
• /
• pp.179-186
• /
• 2007

This paper proposes a novel method for face detection method using Zernike moments. To detect the faces in an image, local regions in multiscale sliding windows are classified into face and non-face by a neural network, and input features of the neural network consist of Zernike moments. Feature dimension is reduced as the reconstruction capability of orthogonal moment. In addition, because the magnitude of Zernike moment is invariant to rotation, a tilted human face can be detected. Even so the detection rate of the proposed method about head on face is less than experiments using intensity features, the result of our method about rotated faces is more robust. If the additional compensation and features are utilized, the proposed scheme may be best suited for the later stage of classification.