• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.36 no.7C
• /
• pp.428-434
• /
• 2011

Recently, subspace analysis has raised its performance to a higher level through the adoption of kernel-based nonlinearity. Especially, the radial basis function, based on its nonparametric nature, has shown promising results in face recognition. However, due to the endemic small sample size problem of face data, the conventional kernel-based feature extraction methods have difficulty in data representation. In this paper, we introduce a novel variant of the RBF kernel to alleviate this problem. By adopting the concept of the nearest feature line classifier, we show both effectiveness and generalizability of the proposed method, particularly regarding the small sample size issue.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.40 no.1
• /
• pp.19-29
• /
• 2003

This paper presents a novel hybrid classifier for face recognition with artificially generated virtual training samples. We utilize both the nearest neighbor approach in feature angle space and a connectionist model to obtain a synergy effect by combining the results of two heterogeneous classifiers. First, a classifier called the nearest feature angle (NFA), based on angular information, finds the most similar feature to the query from a given training set. Second, a classifier has been developed based on the recall of stored frontal projection of the query feature. It uses a frontal recall network (FRN) that finds the most similar frontal one among the stored frontal feature set. For FRN, we used an ensemble neural network consisting of multiple multiplayer perceptrons (MLPs), each of which is trained independently to enhance generalization capability. Further, both classifiers used the virtual training set generated adaptively, according to the spatial distribution of each person's training samples. Finally, the results of the two classifiers are combined to comprise the best matching class, and a corresponding similarit measure is used to make the final decision. The proposed classifier achieved an average classification rate of 96.33% against a large group of different test sets of images, and its average error rate is 61.5% that of the nearest feature line (NFL) method, and achieves a more robust classification performance.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.6 no.6
• /
• pp.986-991
• /
• 2002

This paper presents off-line handwritten numeral recognition method by using Eigen-Vectors. In this method, numeral features are extracted statistically by using Eigen-Vectors through KL transform and input numeral is recognized in the feature space by the nearest-neighbor classifier. In our feature extraction method, basis vectors which express best the property of each numeral type within the extensive database of sample numeral images are calculated, and the numeral features are obtained by using this basis vectors. Through the experiments with the unconstrained handwritten numeral database of Concordia University, we have achieved a recognition rate of 96.2%.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2002.11a
• /
• pp.912-915
• /
• 2002

This paper presents off-line handwritten numeral recognition method by using Eigen-Vectors. In this method, numeral features are extracted statistically by using Eigen-Vectors through KL transform and input numeral is recognized in the feature space by the nearest-neighbor classifier. In our feature extraction method, basis vectors which express best the property of each numeral type within the extensive database of sample numeral images are calculated, and the numeral features are obtained by using this basis vectors. Through the experiments with the unconstrained handwritten numeral database of Concordia University, we have achieved a recognition rate of 96.2%.

• MALSORI
• /
• no.46
• /
• pp.37-50
• /
• 2003

In this paper, we evaluate and compare the performance of speech/music discrimination based on various feature parameters and classifiers. As for feature parameters, we consider High Zero Crossing Rate Ratio (HZCRR), Low Short Time Energy Ratio (LSTER), Spectral Flux (SF), Line Spectral Pair (LSP) distance, entropy and dynamism. We also examine three classifiers: k Nearest Neighbor (k-NN), Gaussian Mixure Model (GMM), and Hidden Markov Model (HMM). According to our experiments, LSP distance and phoneme-recognizer-based feature set (entropy and dunamism) show good performance, while performance differences due to different classifiers are not significant. When all the six feature parameters are employed, average speech/music discrimination accuracy up to 96.6% is achieved.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2011.05a
• /
• pp.185-185
• /
• 2011

Mapping of artificial levee lines, one of major tasks in river zone mapping, is critical to prevention of river flood, protection of environments and eco systems in river zones. Thus, mapping of artificial levee lines is essential for management and development of river zones. Coastal mapping including river zone mapping has been historically carried out using surveying technologies. Photogrammetry, one of the surveying technologies, is recently used technology for national river zone mapping in Korea. Airborne laser scanning has been used in most advanced countries for coastal mapping due to its ability to penetrate shallow water and its high vertical accuracy. Due to these advantages, use of LiDAR data in coastal mapping is efficient for monitoring and predicting significant topographic change in river zones. This paper introduces a method for construction of a 3D artificial levee line using a set of LiDAR points that uses normal vectors. Multiple steps are involved in this method. First, a 2.5-dimensional Delaunay triangle mesh is generated based on three nearest-neighbor points in the LiDAR data. Second, a median filtering is applied to minimize noise. Third, edge selection algorithms are applied to extract break edges from a Delaunay triangle mesh using two normal vectors. In this research, two methods for edge selection algorithms using hypothesis testing are used to extract break edges. Fourth, intersection edges which are extracted using both methods at the same range are selected as the intersection edge group. Fifth, among intersection edge group, some linear feature edges which are not suitable to compose a levee line are removed as much as possible considering vertical distance, slope and connectivity of an edge. Sixth, with all line segments which are suitable to constitute a levee line, one river levee line segment is connected to another river levee line segment with the end points of both river levee line segments located nearest horizontally and vertically to each other. After linkage of all the river levee line segments, the initial river levee line is generated. Since the initial river levee line consists of the LiDAR points, the pattern of the initial river levee line is being zigzag along the river levee. Thus, for the last step, a algorithm for smoothing the initial river levee line is applied to fit the initial river levee line into the reference line, and the final 3D river levee line is constructed. After the algorithm is completed, the proposed algorithm is applied to construct the 3D river levee line in Zng-San levee nearby Ham-Ahn Bo in Nak-Dong river. Statistical results show that the constructed river levee line generated using a proposed method has high accuracy in comparison to the ground truth. This paper shows that use of LiDAR data for construction of the 3D river levee line for river zone mapping is useful and efficient; and, as a result, it can be replaced with ground surveying method for construction of the 3D river levee line.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.10 no.5
• /
• pp.956-961
• /
• 2006

Keystroke time intervals can be a discriminating feature in the verification and identification of computer users. This paper presents a comparison result obtained using several classification methods including k-NN (k-Nearest Neighbor), back-propagation neural networks, and Bayesian classification for keystroke identification. Performance of k-NN classification was best with small data samples available per user, while Bayesian classification was the most superior to others with large data samples per user. Thus, for web-based on-line identification of users, it seems to be appropriate to selectively use either k-NN or Bayesian method according to the number of keystroke samples accumulated by each user.

• Journal of the Korea Computer Graphics Society
• /
• v.20 no.2
• /
• pp.13-24
• /
• 2014

In this study, a graphics-processing-unit (GPU)-based acceleration technique is proposed for the feature-based image morphing. This technique uses the depth-buffer of the graphics hardware to calculate efficiently the shortest distance between a pixel and the control lines. The pairs of control lines between the source image and the destination image are determined by user's input, and the distance function of each control line is rendered using two rectangles and two cones. The distance between each pixel and its nearest control line is stored in the depth buffer through the graphics pipeline, and this is used to conduct the morphing operation efficiently. The pixel-unit morphing operation is parallelized using the compute unified device architecture (CUDA) to reduce the morphing time. We demonstrate the efficiency of the proposed technique using several experimental results.

• Spatial Information Research
• /
• v.6 no.1
• /
• pp.11-23
• /
• 1998

Map generalization functions to visualize the spatial data or to change their scale by changing the level of details of data. Until recently, the studies on map generalization have concentrated more on line features than on point features. However, point features are one of the essential components of digital maps and cannnot be ignored because of the great amount of information they carry. This study, therefore, aimed to find out a detailed procedure of point features' generalization. Particularly, this work chose the distribution pattern of point features as the most important factor in the point generalization in investigating the geometric characteristics of source data. First, it attempted to find out the characteristics of distribution pattern of point features through quadrat analysis with Grieg-Smith method and nearest-neighbour analysis. It then generalized point features through the generalization threshold which did not alter the characteristics of distribution pattern and the removal of redudant point feautres. Therefore, the generalization procedure of point features provided by this work maintained the geometric characteristics as much as possible.

• Proceedings of the KIEE Conference
• /
• 2005.10b
• /
• pp.261-263
• /
• 2005

In this paper, we implement an identification system using the vein image of the hand. The vein pattern is obtained in the grey-scale 2D image through the infrared-red imaging from back of the hand. Since the frame has lack of clearance, we use some enhancing methods such as the complement, addition, and multiplication to the image to increase the contrast. After Wiener filtering for smoothness of the vein pattern, we transform the image into the binary image with mean function. The binarized image is session thinned and the cross-points in the vein tree are obtained by calculating the number of pixels connected because the image is shaped as a tree. We choose the point and find the nearest to the center if it has majority, where we find the two end points of the selected line. We can get the angle between the two lines joined at the cross-point and store its coordinates, angle, and label the values. The values are used as the feature vector of the vein pattern. This procedure is similar to the human cognition sequences. It is shown that the proposed method is simple for the vein recognition.