• Journal of Internet Computing and Services
• /
• v.6 no.6
• /
• pp.187-197
• /
• 2005

In this paper, we propose a method to estimate camera motion parameter based on invariant point features, Typically, feature information of image has drawbacks, it is variable to camera viewpoint, and therefore information quantity increases after time, The LM(Levenberg-Marquardt) method using nonlinear minimum square evaluation for camera extrinsic parameter estimation also has a weak point, which has different iteration number for approaching the minimal point according to the initial values and convergence time increases if the process run into a local minimum, In order to complement these shortfalls, we, first proposed constructing feature models using invariant vector of geometry, Secondly, we proposed a two-stage calculation method to improve accuracy and convergence by using 2D homography and LM method, In the experiment, we compared and analyzed the proposed method with existing method to demonstrate the superiority of the proposed algorithms.

• Journal of Digital Contents Society
• /
• v.2 no.2
• /
• pp.179-189
• /
• 2001

This paper aims at faster processing and retrieval in fingerprint identification systems by reducing the amount of preprocessing and the size of the feature vector. The distribution of fingerprint directions is a set of local directions of ridges and furrows in small overlapped blocks in a fingerprint image. It is extracted initially as a set of 8-direction components through the Gabor filter bank. The discontinuous distribution of directions is smoothed to a continuous one and visualized as a direction image. Then the center of the distribution is selected as a reference point. A feature vector is composed of 192 sine values of the ridge angles at 32-equiangular positions with 6 different distances from the reference point in the direction image. Experiments show that the proposed algorithm performs the same level of correct identification as a conventional algorithm does, while speeding up the overall processing significantly by reducing the length of the feature vector.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.47 no.1
• /
• pp.160-170
• /
• 2010

This paper proposes a Haar Wavelet Feature Detector (HWFD) based on the Haar wavelet transform and average box filter. By decomposing the original image using the Haar wavelet transform, the proposed detector obtains the variance information of the image, making it possible to extract more distinctive features from the original image. For detection of interest points that represent the regions whose variance is the highest among their neighbor regions, we apply the average box filter to evaluate the local variance information and use the integral image technique for fast computation. Due to utilization of the Haar wavelet transform and the average box filter, the proposed detector is robust to illumination change, scale change, and rotation of the image. Experimental results show that even though the proposed method detects fewer interest points, it achieves higher repeatability, higher efficiency and higher matching accuracy compared with the DoG detector and Harris corner detector.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.42 no.2 s.302
• /
• pp.121-130
• /
• 2005

In this paper, we propose a feature extraction method for iris recognition using wavelet transforms. The wavelet transform is fast and has a good localization characteristic. In particular, the low frequency band can be used as an effective feature vector. In iris recognition, the noise caused by eyelid the eyebrow, glint, etc may be included in iris. The iris pattern is distorted by noises by itself, and a feature extraction algorithm based on filter such as Wavelets, Gabor transform spreads noises into whole iris region. Namely, such noises degrade the performance of iris recognition systems a major problem. This kind of noise has adverse effect on performance. In order to solve these problems, we propose to divide the iris image into a number of sub-region and apply the wavelet transform to each sub-region. Experimental results show that the performance of proposed method is comparable to existing methods using Gabor transform and region division noticeably improves recognition performance. However, it is noted that the processing time of the wavelet transform is much faster than that of the existing methods.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.37 no.7C
• /
• pp.554-564
• /
• 2012

The vehicle recognition consists of two steps; the vehicle region detection step and the vehicle identification step based on the feature extracted from the detected region. Features using linear transformations have the effect of dimension reduction as well as represent statistical characteristics, and show the robustness in translation and rotation of objects. Among the linear transformations, the NMF(Non-negative Matrix Factorization) is one of part-based representation. Therefore, we can extract NMF features with sparsity and improve the vehicle recognition rate by the representation of local features of a car as a basis vector. In this paper, we propose a feature extraction using NMF suitable for the vehicle recognition, and verify the recognition rate with it. Also, we compared the vehicle recognition rate for the occluded area using the SNMF(sparse NMF) which has basis vectors with constraint and LVQ2 neural network. We showed that the feature through the proposed NMF is robust in the urban scene where occlusions are frequently occur.

• Journal of the Korea Society of Computer and Information
• /
• v.11 no.1 s.39
• /
• pp.139-145
• /
• 2006

Feature extraction technique in digital images has many applications such as robot vision, medical diagnostic system, and motion video transmission, etc. There are several methods for extracting features in digital images for example nonlinear gradient, nonlinear laplacian, and entropy convolutional filter. However, conventional convolutional filters are usually not efficient to extract features in an image because image feature formation in eyes is more sensitive to dark regions than to bright regions. A few nonlinear filters using difference between arithmetic mean and harmonic mean in a window for extracting sketch features are described in this paper They have some advantages, for example simple computation, dependence on local intensities and less sensitive to small intensity changes in very dark regions. Experimental results demonstrate more successful features extraction than other conventional filters over a wide variety of intensity variations.

• The Bulletin of The Korean Astronomical Society
• /
• v.37 no.2
• /
• pp.99.2-99.2
• /
• 2012

We updated the far-ultraviolet (FUV) spectral images of the entire Vela supernova remnant (SNR) using newly processed FIMS/SPEAR data. In the present study, we compare the newly produced FUV images with the X-ray and $H{\alpha}$ images, and examine how the Vela SNR evolves and interacts with the ambient medium on a global scale. The comparison with X-ray images has revealed a FUV filamentary feature corresponding with the boundary of the northeast-southwest asymmetry of the X-ray shell. The relatively low O IV] ${\lambda}1404$ to O III] ${\lambda}{\lambda}1661$, 1666 ratio estimated on the FUV filament is compatible with the previous proposal that the observed asymmetry of the Vela SNR could be due to the ${\gamma}2$ Velorum stellar wind bubble (SWB). The southwest FUV features surrounding a faint extended X-ray region are characterized as the region where the Vela SNR is interacting slightly stronger with ambient mediums within the dim X-ray southwest section. From a comparison with the $H{\alpha}$ image, we identify a ring-like $H{\alpha}$ feature overlapped with an extended hot X-ray feature of similar size and two local peaks of C IV ${\lambda}{\lambda}1548$, 1551 emission. Their morphologies are consistent with the expected shape when the $H{\alpha}$ ring is in direct contact with the near or far side of the Vela SNR. We suggest that the B3V-type star HD 76161 found at the center of the $H{\alpha}$ ring would be the exciting source of the H II region.

• The Bulletin of The Korean Astronomical Society
• /
• v.37 no.2
• /
• pp.125.1-125.1
• /
• 2012

The wave number 4 (wave-4) and wave number 3 (wave-3) longitudinal structures in the thermospheric neutral mass density are understood as tidal structures driven by diurnal eastward-propagating zonal wave number 3 (DE3) and wave number 2 (DE2) tides, respectively. However, those structures have been identified using data from limited time periods, and the consistency and recurrence of those structures have not yet been examined using long-term observation data. We examine the persistence of those structures by analyzing the neutral mass density data for the years 2001-2008 taken by the CHAllenging Minisatellite Payload (CHAMP) satellite. During years of low solar activity, the amplitude of the wave-4 structure is pronounced during August and September, and the wave-4 phase shows a consistent eastward phase progression of $90^{\circ}$ within 24 h local time in different months and years. During years of high solar activity, the wave-4 amplitude is small and does not show a distinctive annual pattern, but the tendency of the eastward phase shift at a rate of $90^{\circ}$/24 h exists. Thus the DE3 signature in the wave-4 structure is considered as a persistent feature. The wave-3 structure is a weak feature in most months and years. The amplitude and phase of the wave-3 structure do not show a notable solar cycle dependence. Among the contributing tidal modes to the wave-3 structure, the DE2 amplitude is most pronounced. This result may suggest that the DE2 signature, although it is a weak signature, is a perceivable persistent feature in the thermosphere.

• Journal of KIISE:Software and Applications
• /
• v.31 no.5
• /
• pp.605-616
• /
• 2004

A new practical implementation of a facial verification system using cumulative histogram of surface curvatures for the local and contour line areas is proposed, in this paper. The approach works by finding the nose tip that has a protrusion shape on the face. In feature recognition of 3D face images, one has to take into consideration the orientated frontal posture to normalize after extracting face area from the original image. The feature vectors are extracted by using the cumulative histogram which is calculated from the curvature of surface for the contour line areas: 20, 30 and 40, and nose, mouth and eyes regions, which has depth and surface characteristic information. The L1 measure for comparing two feature vectors were used, because it was simple and robust. In the experimental results, the maximum curvature achieved recognition rate of 96% among the proposed methods.

• KIPS Transactions on Software and Data Engineering
• /
• v.3 no.2
• /
• pp.93-100
• /
• 2014

3D object recognition using only 2D images is a difficult work because each images are generated different to according to the view direction of cameras. Because SIFT algorithm defines the local features of the projected images, recognition result is particularly limited in case of input images with strong perspective transformation. In this paper, we propose the object recognition method that improves SIFT algorithm by using several sequential images captured from rotating 3D object around a rotation axis. We use the geometric relationship between adjacent images and merge several images into a generated feature space during recognizing object. To clarify effectiveness of the proposed algorithm, we keep constantly the camera position and illumination conditions. This method can recognize the appearance of 3D objects that previous approach can not recognize with usually SIFT algorithm.