• IEIE Transactions on Smart Processing and Computing
• /
• v.2 no.6
• /
• pp.323-331
• /
• 2013

An extreme learning machine (ELM) is an efficient learning algorithm that is based on the generalized single, hidden-layer feed-forward networks (SLFNs), which perform well in classification applications. Many studies have demonstrated its superiority over the existing classical algorithms: support vector machine (SVM) and BP neural network. This paper presents a novel face recognition approach based on a multi-class project extreme learning machine (MPELM) classifier and 2D Gabor transform. First, all face image features were extracted using 2D Gabor filters, and the MPELM classifier was used to determine the final face classification. Two well-known face databases (CMU-PIE and ORL) were used to evaluate the performance. The experimental results showed that the MPELM-based method outperformed the ELM-based method as well as other methods.

• Journal of Institute of Control, Robotics and Systems
• /
• v.14 no.8
• /
• pp.818-823
• /
• 2008

In this paper, we propose a feature extraction method which extracts directional features of handwritten numerals by using the projection runlength. Our directional featrures are obtained from four directional images, each of which contains horizontal, vertical, right-diagonal and left-diagonal lines in entire numeral shape respectively. A conventional method which extracts directional features by using Kirsch masks generates edge-shaped double line directional images for four directions, whereas our method uses the projections and their runlengths for four directions to produces single line directional images for four directions. To obtain the directional projections for four directions from a numeral image, some preprocessing steps such as thinning and dilation are required, but the shapes of resultant directional lines are more similar to the numeral lines of input numerals. Four [$4{\times}4$] directional features of a numeral are obtained from four directional line images through a zoning method. By using a hybrid feature which is made by combining our feature with the conventional features of a mesh features, a kirsch directional feature and a concavity feature, higher recognition rates of the handwrittern numerals can be obtained. For recognition test with given features, we use a multi-layer perceptron neural network classifier which is trained with the back propagation algorithm. Through the experiments with the handwritten numeral database of Concordia University, we have achieved a recognition rate of 97.85%.

• Management Science and Financial Engineering
• /
• v.7 no.1
• /
• pp.27-39
• /
• 2001

The multi-projective model considers attributes and the relationships among attributes called projections. The critical features of the multi-projective model are the way of relating attributes in the description of the system, the way of reasoning incomplete projections, and the determination of connected patterns between projection. In order to get a full picture of the system, we build a set of projections. The multi-projective model can be thought of as projections of a multi-dimensional reality onto simplified “model space”. The multi-projective database modeling approach used in this paper unified the ideas and terminology of various database models. Most importantly, the multi-projective modeling is presented as a tool of database design in the relational and other database models.

• Journal of the Korea Computer Graphics Society
• /
• v.21 no.1
• /
• pp.1-11
• /
• 2015

Researchers have developed virtual reality environments to provide audience with more visually immersive experiences than previously possible. One of the most popular solutions to build the immersive VR space is a multi-projection technique. However, utilization of multiple projectors requires large spaces, expensive cost, and accurate geometry calibration among projectors. This paper presents a novel omnidirectional projection system with a single projector and a single spherical mirror.We newly designed the simple and intuitive calibration system to define the shape of environment and the relative position of mirror/projector. For successful image projection, our optimized omnidirectional image generation step solves image distortion produced by the spherical mirror and a calibration problem produced by unknown parameters such as the shape of environment and the relative position between the mirror and the projector. Additionally, the focus correction is performed to improve the quality of the projection. The experiment results show that our method can generate the optimized image given a normal panoramic image for omnidirectional projection in a rectangular space.

• The Journal of the Korea Contents Association
• /
• v.7 no.2
• /
• pp.84-91
• /
• 2007

Recently, the researches based on vision about user attention and action awareness are being pushed actively for human computer interaction. Among them, various applications of tabletop display system are developed more in accordance with touch sensing technique, co-located and collaborative work. Formerly, although supported only one user, support multi-user at present. Therefore, collaborative work and interaction of four elements (human, computer, displayed objects, physical objects) that is ultimate goal of tabletop display are realizable. Generally, tabletop display system designs according to four key aspects. 1)multi-touch interaction using bare hands. 2)implementation of collaborative work, simultaneous user interaction. 3)direct touch interaction. 4)use of physical objects as an interaction tool. In this paper, we describe a critical analysis of the state-of-the-art in advanced multi-touch sensing techniques for tabletop display system according to the four methods: vision based method, non-vision based method, top-down projection system and rear projection system. And we also discuss some problems and practical applications in the research field.

• Proceedings of the Acoustical Society of Korea Conference
• /
• autumn
• /
• pp.261-264
• /
• 2004

본 논문에서 는 muti-channel lattice 예측기를 사용하여 AP(affine projection) 알고리듬을 근사적으로 구현하는 알고리즘을 제안한다. Lattice 예측기의 예측 오차를 사용하여 TDL 필터 계수를 적응적으로 조정함으로써 AP 알고리듬을 근사화한다. 또한 전처리단으로 사용된 lattice 예측기를 TDL 필터와 결합함으로써 기존의 방법보다 계산량을 더욱 줄일 수 있는 알고리듬을 제안한다. 제안된 알고리즘은 기존의 알고리즘보다 적은 계산량을 필요로 하지 만 AP 알고리즘을 보다 근사적으로 구현할 수 있다는 장점이 있다.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2008.11a
• /
• pp.665-666
• /
• 2008

• Proceedings of the KWS Conference
• /
• 2005.11a
• /
• pp.105-107
• /
• 2005

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.3C
• /
• pp.419-425
• /
• 2004

This paper presents a new method which is calculated to use only three vanishing points in order to compute the dimensions of object and its pose from a single image of perspective projection taken by a camera. Our approach is to only compute three vanishing points without informations such as the focal length and rotation matrix from images in the case of perspective projection. We assume that the object can be modeled as a linear function of a dimension vector v. The input of reconstruction is a set of correspondences between features in the model and features in the image. To minimize each the dimensions of the parameterized models, this reconstruction of optimization can be solved by standard nonlinear optimization techniques with a multi-start method which generates multiple starting points for the optimizer by sampling the parameter space uniformly.

• 한국HCI학회:학술대회논문집
• /
• 2006.02a
• /
• pp.1081-1086
• /
• 2006

일반적으로 프로젝터는 심도(depth of field)가 제한되어 있기 때문에 스크린이 곡면일 경우, 일부 영역에서는 초점이 맞지 않게(out-of-focus) 된다. 이런 out-of-focus 영역의 정보는 블러링(blurring)되기 때문에 사용자에게 정확한 정보를 전달할 수 없다. 여러 대의 프로젝터를 이용할 경우, 각 프로젝터는 다른 in-focus 영역을 가지기 때문에 각 프로젝터 픽셀의 in-focus/out-of-focus 판별을 통해 in-focus 픽셀만을 투사함으로써, out-of-focus 픽셀의 영향을 제거할 수 있다. 그러나 여러 대의 프로젝터의 in-focus영역이 거의 일치할 경우, out-of-focus영역은 여전히 out-of-focus상태일 수 밖에 없다. 따라서, 각 프로젝터의 초점을 유연하게 조절하면서 동시에 여러 대의 프로젝터에 의한 in-focus 영역을 최대한 크게 할 수 있는 방법이 필요하다. 본 논문에서는 각 프로젝터의 초점을 유동적으로 조절하면서 취득된 영상을 처리하여 in-focus 영역을 판별하고 각 프로젝터에 대한 in-focus 영역을 조합하여 전체in-focus 영역의 면적을 최대화하는 방법을 제안한다. 제안된 방법의 유용성을 검증하기 위해, 각 프로젝터의 in-focus 영역을 적절한 컬러를 이용해 시각적으로 표현해주고 이 정보를 참조해 각 프로젝터의 초점을 유동적으로 조절하여 전초점 영상을 만들어내는 시스템을 구현하였다.