• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.64 no.6
• /
• pp.900-911
• /
• 2015

In this paper, 3D face recognition model is designed by using Polynomial based RBFNN(Radial Basis Function Neural Network) and PNN(Polynomial Neural Network). Also recognition rate is performed by this model. In existing 2D face recognition model, the degradation of recognition rate may occur in external environments such as face features using a brightness of the video. So 3D face recognition is performed by using 3D scanner for improving disadvantage of 2D face recognition. In the preprocessing part, obtained 3D face images for the variation of each pose are changed as front image by using pose compensation. The depth data of face image shape is extracted by using Multiple point signature. And whole area of face depth information is obtained by using the tip of a nose as a reference point. Parameter optimization is carried out with the aid of both ABC(Artificial Bee Colony) and PSO(Particle Swarm Optimization) for effective training and recognition. Experimental data for face recognition is built up by the face images of students and researchers in IC&CI Lab of Suwon University. By using the images of 3D face extracted in IC&CI Lab. the performance of 3D face recognition is evaluated and compared according to two types of models as well as point signature method based on two kinds of depth data information.

• Korean Journal of Computational Design and Engineering
• /
• v.13 no.6
• /
• pp.403-411
• /
• 2008

The alignment of facial images is crucial for 2D face recognition. This is the same to facial meshes for 3D face recognition. Most of the 3D face recognition methods refer to 3D alignment but do not describe their approaches in details. In this paper, we focus on describing an automatic 3D alignment in viewpoint of quantitative analysis. This paper presents a framework of 3D face alignment and normalization based on feature points obtained by Active Shape Models (ASMs). The positions of eyes and mouth can give possibility of aligning the 3D face exactly in three-dimension space. The rotational transform on each axis is defined with respect to the reference position. In aligning process, the rotational transform converts an input 3D faces with large pose variations to the reference frontal view. The part of face is flopped from the aligned face using the sphere region centered at the nose tip of 3D face. The cropped face is shifted and brought into the frame with specified size for normalizing. Subsequently, the interpolation is carried to the face for sampling at equal interval and filling holes. The color interpolation is also carried at the same interval. The outputs are normalized 2D and 3D face which can be used for face recognition. Finally, we carry two sets of experiments to measure aligning errors and evaluate the performance of suggested process.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.20 no.1
• /
• pp.116-122
• /
• 2010

This paper proposes a method for classifying targets robust to geometric transformations of targets such as rotation, scale change, translation, and pose change. Targets which have rotation, scale change, and shift is firstly classified based on CM(Confidence Map) which is generated by similarity, scale ratio, and range of orientation for SIFT(Scale-Invariant Feature Transform) feature vectors. On the other hand, DB(DataBase) which is acquired in various angles is used to deal with pose variation of targets. Range of the angle is determined by comparing and analyzing the execution time and performance for sampling intervals. We experiment on various images which is geometrically changed to evaluate performance of proposed target classification method. Experimental results show that the proposed algorithm has a good classification performance.

• Journal of Korean Society of Occupational and Environmental Hygiene
• /
• v.20 no.4
• /
• pp.217-224
• /
• 2010

The purity methods to determine LOD/LOQ using standard deviation of the residual, intercept and blank by IUPAC and ACS describe many of the pitfalls and pose significant challenges to analytical chemists. Therefore, the aim of this study is the development of the simple, easy, convenient and statistically significant method to determine LOD in quantitative analysis of organic solvents by GC. The new approaching method by linearization in the given concentration range used coefficient of variation ; ${\sigma}_{n-1}$/S(standard deviation, ${\sigma}_{n-1}$ and average, S) of sensitivity(Response/concentration). The comparison of results among the purity methods(IUPAC and ACS) and the linearization have been fulfilled the F-test for standard deviations and t-test for LOD range values. The results of F-test and t-test are satisfied within 95 % confidence level, respectably. The LOD values determined by the new procedure are n-Hexane 0.0116 mg/$m^3$, Toluene 0.0807 mg/$m^3$, and o-Xylene 0.0494 mg/$m^3$. Because the standard deviation of the residual, intercept and blank and the slope of calibration curve are not calculated and the new approaching method use the coefficient of variation of sensitivity by linearization, this new method is simple, easy, convenient and statistically significant. In future, many chemical analysts will expect to applicate and routinely use this method in the all quantitative analysis.

• Proceedings of the KIEE Conference
• /
• 2004.11c
• /
• pp.30-32
• /
• 2004

With 3-D vision measuring, camera calibration is necessary to calculate parameters accurately. Camera calibration was developed widely in two categories. The first establishes reference points in space, and the second uses a grid type frame and statistical method. But, the former has difficulty to setup reference points and the latter has low accuracy. In this paper we present an algorithm for camera calibration using perspective ratio of the grid type frame with different line widths. It can easily estimate camera calibration parameters such as lens distortion, focal length, scale factor, pose, orientations, and distance. The advantage of this algorithm is that it can estimate the distance of the object. Also, the proposed camera calibration method is possible estimate distance in dynamic environment such as autonomous navigation. To validate proposed method, we set up the experiments with a frame on rotator at a distance of 1, 2, 3, 4[m] from camera and rotate the frame from -60 to 60 degrees. Both computer simulation and real data have been used to test the proposed method and very good results have been obtained. We have investigated the distance error affected by scale factor or different line widths and experimentally found an average scale factor that includes the least distance error with each image. The average scale factor tends to fluctuate with small variation and makes distance error decrease. Compared with classical methods that use stereo camera or two or three orthogonal planes, the proposed method is easy to use and flexible. It advances camera calibration one more step from static environments to real world such as autonomous land vehicle use.

• Journal of Korea Multimedia Society
• /
• v.15 no.1
• /
• pp.63-70
• /
• 2012

In this paper an extension to an original active shape model(ASM) for facial feature extraction is presented. The original ASM suffers from poor shape alignment by aligning the shape model to a new instant of the object in a given image using a simple similarity transformation. It exploits only informations such as scale, rotation and shift in horizontal and vertical directions, which does not cope effectively with the complex pose variation. To solve the problem, new shape alignment with 6 degrees of freedom is derived, which corresponds to an affine transformation. Another extension is to speed up the calculation of the Mahalanobis distance for 2-D profiles by trimming the profile covariance matrices. Extensive experiment is conducted with several images of varying poses to check the performance of the proposed method to segment the human faces.

• 제어로봇시스템학회:학술대회논문집
• /
• 2004.08a
• /
• pp.1048-1052
• /
• 2004

In this work, we have investigated a role of linking parameter in Pulse-Coupled Neural Network(PCNN) which is suggested to explain the synchronous activities among neurons in the cat cortex. Then we have found a method to determine the linking parameter for a satisfactory face detection performance in a given color image. Face detection algorithm which uses the color information is independent on pose, size and obstruction of a face. But the use of color information encounters some problems arising from skin-tone color in the background, intensity variation within faces, and presence of random noise and so on. Depending on these conditions, PCNN's linking parameters should be selected an appropriate values. First we obtained the mean and variance of the skin-tone colors by experiments. Then, we introduced a preprocess that the pixel with a mean value of skin-tone colors has the highest level value (255) and the other pixels have values between 0 and 255 according to normal distribution with a variance. This preprocessing leads to an easy decision of the linking parameter of the Pulse-Coupled Neural Network. Through experiments, it is verified that the proposed method can improve the face detection performance compared to the existing methods.

• Journal of KIISE:Software and Applications
• /
• v.31 no.6
• /
• pp.784-791
• /
• 2004

In this study, we deal with face recognition using fuzzy-based Fisherface method. The well-known Fisherface method is more insensitive to large variation in light direction, face pose, and facial expression than Principal Component Analysis method. Usually, the various methods of face recognition including Fisherface method give equal importance in determining the face to be recognized, regardless of typicalness. The main point here is that the proposed method assigns a feature vector transformed by PCA to fuzzy membership rather than assigning the vector to particular class. In this method, fuzzy membership degrees are obtained from FKNN(Fuzzy K-Nearest Neighbor) initialization. Experimental results show better recognition performance than other methods for ORL and Yale face databases.

• Proceedings of the IEEK Conference
• /
• 2003.07e
• /
• pp.1968-1971
• /
• 2003

We propose a knowledge-based algorithm for extracting an object boundary from low-quality image like the forward looking infrared image. With the multi-classes training data set, the global shape is modeled by multispace KL(MKL)[1] and curvature model. And the objective function for fitting the deformable boundary template represented by the shape model to true boundary in an input image is formulated by Bales rule. Simulation results show that our method has more accurateness in case of multi-classes training set and performs better in the sense of computation cost than point distribution model(PDM)[2]. It works well in distortion under the noise, pose variation and some kinds of occlusions.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2002.10a
• /
• pp.355-360
• /
• 2002

To calibrate a non-redundantly actuated parallel mechanism, one can find actual kinematic parameters by means of geometrical constraint of the mechanism's kinematic structure and measurement values. However, the calibration algorithm for a non-redundant case does not apply fur a redundantly actuated parallel mechanism, because the angle error of the actuating joint varies with position and the geometrical constraint fails to be consistent. Such change of joint angle error comes from constraint torque variation with each kinematic pose (meaning position and orientation). To calibrate a redundant parallel mechanism, one therefore has to consider constraint torque equilibrium and the relationship of constraint torque to torsional deflection, in addition to geometric constraint. In this paper, we develop the calibration algorithm fir a redundantly actuated parallel mechanism using these three relationships, and formulate cost functions for an optimization algorithm. As a case study, we executed the calibration of a 2-DOF parallel mechanism using the developed algorithm. Coordinate values of tool plate were measured using a laser ball bar and the actual kinematic parameters were identified with a new cost function of the optimization algorithm. Experimental results showed that the accuracy of the tool plate improved by 82% after kinematic calibration in a redundant actuation case.