Adaptive Image Segmentation for Facial Photos

Title Adaptive Image Segmentation for Facial Photos
Journal CISC-S2008 pp.269-273
Author  Hale Kim
Abstract This paper presents an adaptive image segmentation for facial photos that is independent to background colors and robust to shadows on backgrounds. Firstly, a coarse segmentation is obtained by combining the Canny edge detector and morphological operations. Then, the input image is converted into normalized RGB color model and each normalized pixel is then classified into the foreground or background. Finally, detection of face candidate regions is done using skin color information. The combination of the face candidate regions and the segmentation in normalized color model is carried out to achieve foreground segmentation without shadows.
Keywords

병렬처리를 이용한 특징점 추출 알고리즘의 고속화

Title 병렬처리를 이용한 특징점 추출 알고리즘의 고속화
Journal 제어자동화시스템 심포지엄 (CASS 2008) pp.613-618
Author 박은수, 최학남, 김학일
Abstract In this paper, we present a faster more efficient object recognition method for moving robots. Recognizing places or objects in real time can be a difficult task due to many different features. To attain this task, we implemented a Parallel feature extract detector using OpenMP. In the first step, Algorithms and codes are optimized in order to execute parallel processing. The processed algorithms in parallel are debugged for robust Performance. As a result, the intelligent conversion of the source codes to exploit OpenMP technology increase processing speed by two and half times over previous detectors while still maintaining robust performance measured using a standard evaluation set
Keywords

GPU를 이용한 고속 영상 보간법 개발

Title GPU를 이용한 고속 영상 보간법 개발
Journal CICS 2008 pp.300-301
Author 최학남, 박은수,  김학일
Abstract 본 논문에서는 GPU를 이용한 고속 보간법 개발방법에 대해 제안한다. GPU는 흔히 그래픽 연산에 사용되지만 최근에는 GPGPU가 각광을 받고 있다. 특히 NVIDIA에서 발표한 CUDA를 이용하면 GPU를 쉽게 접근하여 프로세싱 할 수 있어 많은 분야에서 GPU를 활용하고 있다. 본 논문에서는 실제 CUDA를 이용하여 여러 가지 보간법에 대한 알고리즘을 구현하여 CUDA의 성능을 확인하였다. CPU에서 구현한 알고리즘과 CUDA를 이용한 알고리즘을 비교했을 때 메모리 할당 및 전송부분을 제외한 수순 프로세싱 시간을 보면 GPU에서 훨씬 좋은 성능을 나타내었고, 메모리 할당 및 전송을 고려했을 때 작은 사이즈 영상에서는 오히려 역효과가 나타났고, 대용량 영상에서는 좋은 성능을 나타냄을 확인하였다.
Keywords

Comparative Assessment of Fingerprint Sample Quality Measures

Title Comparative Assessment of Fingerprint Sample Quality Measures
Journal CISC-W08-한국정보보호학희 동계정보보호학술대희 논문집pp.228-232
Author Hakil Kim
Abstract Fingerprint sample quality is one of major factors influencing the matching performance of fingerprint recognition systems. The purpose of this paper is to assess the effectiveness of individual sample quality measures on the performance of minutiae-based fingerprint recognition algorithms. Initially, the authors examined the various factors that influenced the matching performance of the minutiae-based fingerprint recognition algorithms. Then, the existing measures for fingerprint sample quality were studied and the more effective quality measures were selected and compared with two image quality software packages in terms of matching performance. The experimental results over various datasets show that even a single sample quality measure can enhance the matching performance effectively.
Keywords

Head Location in Facial Photos Based on Facial Features,

Title Head Location in Facial Photos Based on Facial Features,
Journal CISC-W2008 pp.233-237
Author  Hale Kim
Abstract This paper presents a method to locate the head in facial photos based on skin color and facial features, namely mouth, chin and crown. Firstly, the image segmentation process is performed to separate the subject and the background. Then, skin pixels are extracted from the original image and the direction of the head is estimated based on the orientation of skin pixels. Finally, detections of mouth, chin and crown are done
Keywords

Performance study on methodologies of quality evaluation of fingerprint

Title Performance study on methodologies of quality evaluation of fingerprint
Journal CISC-S08-한국정보보호학희 동계정보보호학술대희 논문집pp.147-152
Author  Hakil Kim
Abstract Fingerprint sample quality evaluation (FSQE) is an important issue in fingerprint recognition. This paper reviews existing methodologies for FSQE and gives out the definition of fingerprint quality. Then, the representative set of quality measures is analyzed by studying their performance and their correlation. The purpose of this work is to produce a single scalar quantity based on human perception by combining existing quality measures. For the experiments, eleven FVC data sets were selected. In order to compare the distribution of fingerprint quality scores, all of the selected databases were combined into one database by ignoring the difference of sensor type and be tested for each quality measures.
Keywords

GPU를 이용한 고속 Texture 특징점 추출 알고리즘 개발

Title GPU를 이용한 고속 Texture 특징점 추출 알고리즘 개발
Journal IPIU 2009 pp.826-830
Author 최학남, 박은수,김학일
Abstract 본 논문에서는 GPU를 이용한 고속 픽셀 패턴 기반의 Texture 특징점 추출 알고리즘을 제안한다. 제안한 알고리즘은 Gabor 필터를 이용하여 영상 내에서 각각의 픽셀에 대한 texture 정보를 추출하고, 추출된 정보를 이용하여 Pattern Map을 구성하며, 마지막으로 Pattern Map을 이용하여 Texture 특징 벡터를 추출한다. 또한 Gabor texture 특징 벡터 추출 과정의 성능을 평가하기 위하여 CPU상에서는 순차처리와 OpenMP를 이용한 병렬처리를 하였고, GPU상에서는 Global 메모리를 이용한 방법과 Shared 메모리를 이용한 방법으로 개발하였다. 실험결과 Shared 메모리를 이용한 GPU상에서의 처리속도가 가장 좋은 성능을 나타내는 것을 확인하였다.
Keywords

실시간 물체 인식을 위한 특징 추출 알고리즘의 병렬화

Title 실시간 물체 인식을 위한 특징 추출 알고리즘의 병렬화
Journal IPIU 2009 pp.772-777
Author 박은수, 최학남, 김학일
Abstract This paper presents a fast feature extraction method for mobile robots by parallel processing based on OpenMP and SSE (Streaming SIMD Extension) programming. Recognizing scenes or objects in real-time is a difficult task, due to the variety of features. To complete this task, a parallel feature extraction detector and descriptor via OpenMP and SSE is proposed. In the first step, the algorithms and codes are optimized, in order to be implemented by parallel processing. The implemented parallel algorithms are debugged, to maintain the same level of performance. The process of extracting key points and obtaining the dominant orientation with respect to key points is parallelized via these steps. After extraction, a parallel descriptor via SSE instructions is constructed. As a result, the proposed Parallel-Hessian accelerates processing by up to two and half times over previous detectors, and the descriptor achieves an acceleration of up to four and half times compared to the original SIFT, while maintaining a robust performance measured via a standard evaluation set
Keywords

Gaussian Filter-Based Directional Fields Estimation in Fingerprint Images

Title Gaussian Filter-Based Directional Fields Estimation in Fingerprint Images
Journal IPIU 2009 pp.341-344
Author Hakil Kim
Abstract Directional Fields play important roles in fingerprint recognition. A robust method of pixel-wise DFs estimation based on Gaussian filter is proposed in this paper. Firstly, the covariance data for each pixel were computed and smoothed by a Gaussian filter. Then the sine and cosine components are computed from the squared gradient vectors which were derived from covariance data, and filtered by the Gaussian filter again. Finally, the pixel-wise DFs were obtained from the inverse tangent function. The FVC databases were employed in this study and the experimental results show the proposed method can work effectively for fingerprints captured from any sensors in the FVC databases.
Keywords

Singular Point Detection from High-resolution Orientation Field

Title Singular Point Detection from High-resolution Orientation Field
Journal CISC-S2009 pp.313-317
Author  Hakil Kim
Abstract A singular point detection method based on high-resolution orientation field is proposed in this paper. Firstly, complex angle differential is computed from the high-resolution Orientation Field (OF), then, the singular point in OF is extracted and taken as candidate singular point. Finally, the Poincare Index (PI) is utilized to verify the candidate point. The proposed algorithm is tested on FVC2000 DB 2b, and shows robust experimental results.
Keywords

Lip Feature Extraction for Facial Image Quality Assessment

Title Lip Feature Extraction for Facial Image Quality Assessment
Journal CISC-S2009 pp.318-322
Author Hale Kim
Abstract This paper proposes a robust method for analyzing color facial images to extract four lip features, namely, mouth corners and the centers of inner upper/lower lips, upon a condition that the positions of two eye centers are provided in advance. Four lip features are located by taking the characteristic of lip color and edge information into account. The strong relation between positions of eyes and mouth is utilized to restrict the area where the lip features are observed. An experimental validation has been conducted on 300 images collected from the FERET database and has demonstrated the accuracy and robustness of the proposed method.
Keywords

SIFT를 기반으로 하는 PCB 영상의 정렬 알고리즘 개발

Title SIFT를 기반으로 하는 PCB 영상의 정렬 알고리즘 개발
Journal 대한전자공학회 하계종합학술대회2009 (IEEK Summer Conference 2009) 제 31권 제 1호, pp.1346-1347
Author  최학남, 김학일
Abstract This paper presents an image alignment algorithm for application of AOI (Automatic optical inspection) based on SIFT. Since the correspondences result using SIFT descriptor have many wrong points for aligning, this paper modified and classified those points by five measures. Experimental results show that the proposed method has similar rotation and robust translation accuracy in comparison to the commercial software MIL 8.0
Keywords

STATISTICAL NOISE BAND REMOVAL FOR SURFACE CLUSTERING OF HYPERSPECTRAL DATA

Title STATISTICAL NOISE BAND REMOVAL FOR SURFACE CLUSTERING OF HYPERSPECTRAL DATA
Journal Proceedings of International Symposium on Remote Sensing pp 111~114
Author  Hakil Kim
Abstract The existence of noise bands may deform the typical shape of the spectrum, making the accuracy of clustering degraded. This paper proposes a statistical approach to remove noise bands in hyperspectral data using the correlation coefficient of bands as an indicator. Considering each band as a random variable, two adjacent signal bands in hyperspectral data are highly correlative. On the contrary, existence of a noise band will produce a low correlation. For clustering, the unsupervised k-nearest neighbor clustering method is implemented in accordance with three well-accepted spectral matching measures, namely ED, SAM and SID. Furthermore, this paper proposes a hierarchical scheme of combining those. Finally, a separability assessment based on the between-class and the within-class scatter matrices is followed to evaluate the applicability of the proposed noise band removal method. Also, the paper brings out a comparison for spectral matching measures. The experimental results conducted on a 228-band hyperspectral data show that while the SAM measure is rather resistant, the performance of SID measure is more sensitive to noise.
Keywords

Location of Iris Based on Circular and Linear Filters

Title Location of Iris Based on Circular and Linear Filters
Journal Proceedings of the 10th International Conference on Control, Automation, Robotics and Vision, ICARCV 2008
Author Hakil Kim
Abstract This paper presents a novel method of iris location for iris recognition. In detecting the inner boundary of the iris, the method introduces a circular filter which is specially designed to detect circular shapes in iris images. The inner boundary detection process consists of two steps. Firstly, a coarse center of the pupil is detected by the proposed filter. Then, a set of inner boundary points is found by applying the linear Hough transform to rectangular areas, converted from extracted regions based on the coarse center. The inner boundary is finally determined by using the least square method to fit the set of points to a circle. The presented method further proposes a process of extracting the outer boundary of the iris by applying a linear filter to the rectangular areas transformed from arc regions around the iris. Experimental results show that the proposed method performs promisingly on iris images captured in ideal and non-ideal conditions from the CASIA iris database.
Keywords

Liveness detection of fingerprint based on band-selective fourier spectrum

Title Liveness detection of fingerprint based on band-selective fourier spectrum
Journal ICISC 2007 – 10th international conference on information security and cryptology pp. 168-179
Author  Hakil Kim
Abstract This paper proposes a novel method for fingerprint liveness detection based on band-selective Fourier spectrum. The 2D spectrum of a fingerprint image reflects the distribution and strength in spatial frequencies of ridge lines. The ridge-valley texture of the fingerprint produces a ring pattern around the center in the Fourier spectral image and a harmonic ring pattern in the subsequent ring. Both live and fake fingerprints produce these rings, but with different amplitudes in different spatial frequency bands. Typically, live fingerprints show stronger Fourier spectrum in the ring patterns than the fake. The proposed method classifies the live and the fake fingerprints by analyzing the band-selective Fourier spectral energies in the two ring patterns. The experimental results demonstrate this approach to be a promising technique for making fingerprint recognition systems more robust against fake-finger-based spoofing vulnerabilities.
Keywords

Comparative Assessment of Fingerprint Sample Quality Measures Based on Minutiae-based Matching Performance

Title Comparative Assessment of Fingerprint Sample Quality Measures Based on Minutiae-based Matching Performance
Journal ISECS – 2nd pp. 309-313
Author  Hakil Kim
Abstract This Fingerprint sample quality is one of major factors influencing the matching performance of fingerprint recognition systems. The error rates of fingerprint recognition systems can be decreased significantly by removing poor quality fingerprints. The purpose of this paper is to assess the effectiveness of individual sample quality measures on the performance of minutiae-based fingerprint recognition algorithms. Initially, the authors examined the various factors that influenced the matching performance of the minutiae-based fingerprint recognition algorithms. Then, the existing measures for fingerprint sample quality were studied and the more effective quality measures were selected and compared with two image quality software packages, (NFIQ from NIST, and QualityCheck from Aware Inc.) in terms of matching performance of a commercial fingerprint matcher (Verifinger 5.0 from Neurotechnologija). The experimental results over various Fingerprint Verification Competition (FVC) datasets show that even a single sample quality measure can enhance the matching performance effectively.
Keywords

A Fast Feature Extraction in Objectr Recognition Using Parallel processing on CPU and GPU

Title A Fast Feature Extraction in Objectr Recognition Using Parallel processing on CPU and GPU
Journal 2009 IEEE International Conference on Systems, Man, and Cybernetics
Author  Eunsoo Park, Xuenan Cui, Hakil Kim
Abstract Due to the advents of multi-core CPU and GPU, various parallel processing techniques have been widely applied to many application fields including computer vision. This paper presents a parallel processing technique for realtime feature extraction in object recognition by autonomous mobile robots, which utilizes both CPU and GPU by combining OpenMP, SSE (Streaming SIMD Extension) and CUDA programming. Firstly, the algorithms and codes for feature extraction are optimized and implemented in parallel processing. After the parallel algorithms are assured to maintain the same level of performance, the process for extracting key points and obtaining dominant orientation with respect to the key points is parallelized. Following the extraction is the construction of a parallel descriptor via SSE instructions. Finally, the GPU version of SIFT is also implemented using CUDA. The experiments have shown that the CPU version of SIFT is almost five times faster than the original SIFT while maintaining robust performance. Further, the GPU-Parallel descriptor achieves acceleration up to five times higher than the CPU-Parallel descriptor at a cost of a bit lower performance.
Keywords

Eye-verifier Using Ternary Template for Reliable Eye Detection in Facial Color Images

Title Eye-verifier Using Ternary Template for Reliable Eye Detection in Facial Color Images
Journal IEEE Third international conference on biometrics: Theory, Application and System (BTAS 09)
Author Hakil Kim
Abstract This paper introduces an eye-verifier for reliable detection of eyes in facial color images. At first, the eye region candidates are searched by the circular filter to the binary face image. The eye candidates are then fed into an eye-verifier. The eye-verifier uses a ternary template, generated from the eye area consisting of iris, sclera and skin. Then the template matching is made by ternary Hamming distance. Experimental results evaluating the proposed method on our own face database and FERET database show promising performance in terms of detection rate. The proposed algorithm showed a robust performance to the face poses and eye directions
Keywords

Eye Feature Extraction Using K-means Clustering for Low Illumination and Iris Color Variety

Title Eye Feature Extraction Using K-means Clustering for Low Illumination and Iris Color Variety
Journal 11th International Conference on Control, Automation, Robotics and Vision (ICARCV 2010)
Author  Hakil Kim
Abstract This paper presents an approach for locating eye features in color images based on the unsupervised K-means clustering. Given the assumption that the input is an eye window containing a single eye, the proposed method detects the iris by unsupervised K-means clustering on the feature spaces of compensated red and green color channels. The iris circle is then refined using the gradient information and circular Hough transform. For the sclera detection, the r-g and r-b are utilized as they show the discriminant feature of sclera regardless of light condition and iris color. The sclera is then extended to fit the eyelids by a region growing scheme. Experiments on a collection of eye images extracted from FERET facial database and our self-collected images show a promising performance toward the low illumination and iris color variety.
Keywords

Noisy Band Removal Using Band Correlation in Hyperspectral Images

Title Noisy Band Removal Using Band Correlation in Hyperspectral Images
Journal Korean Journal of Remote Sensing (대한원격탐사학회 영문저널) vol 25, no. 3 pp.263~270. (2009)
Author Hakil Kim
Abstract Noise band removal is a crucial step before spectral matching since the noise bands can distort the typical shape of spectral reflectance, leading to degradation on the matching results. This paper proposes a statistical noise band removal method for hyperspectral data using the correlation coefficient between two bands. The correlation coefficient measures the strength and direction of a linear relationship between two random variables. Considering each band of the hyperspectral data as a random variable, the correlation between two signal bands is high; existence of a noisy band will produce a low correlation due to ill-correlativeness and undirectedness. The unsupervised k-nearest neighbor clustering method is implemented in accordance with three well-accepted spectral matching measures, namely ED, SAM and SID in order to evaluate the validation of the proposed method. This paper also proposes a hierarchical scheme of combining those measures. Finally, a separability assessment based on the between-class and the within-class scatter matrices is followed to evaluate the applicability of the proposed noise band removal method. Also, the paper brings out a comparison for spectral matching measures. The experimental results conducted on a 228-band hyperspectral data show that while the SAM measure is rather resistant, the performance of SID measure is more sensitive to noise.
Keywords Noise band removal, hierarchical classification, spectral matching measure, separability assessment

Design and Implementation of Hyperspectral Image Analysis Tool: HYVIEW

Title Design and Implementation of Hyperspectral Image Analysis Tool: HYVIEW
Journal Korean Journal of Remote Sensing (대한원격탐사학회 영문저널) vol 23, no. 3 pp.171~179.
Author  Hakil Kim
Abstract Hyperspectral images have shown a great potential for the applications in resource management, agriculture, mineral exploration and environmental monitoring. However, due to the large volume of data, processing of hyperspectral images faces some difficulties. This paper introduces the development of an image processing tool (HYVIEW) that is particularly designed for handling hyperspectral image data. Current version of HYVIEW is dealing with efficient algorithms for displaying hyperspectral images, selecting bands to create color composites, and atmospheric correction. Three band-selection schemes for producing color composites are available based on three most popular indexes of OIF, SI and CI. HYVIEW can effectively demonstrate the differences in the results of the three schemes. For the atmospheric correction, HYVIEW utilizes a pre-calculated LUT by which the complex process of correcting atmospheric effects can be performed fast and efficiently.
Keywords

3축 가속도 센서의 흔들림 정보를 이용한 Deblurring

Title 3축 가속도 센서의 흔들림 정보를 이용한 Deblurring
Journal 대한전자공학회논문지 Vol. 45, No.3 pp.1~11
Author  박은수, 김학일
Abstract 본 논문은 모바일 단말기에 탑재된 카메라를 이용하여 정지영상을 획득할 때 발생할 수 있는 blur현상을 3축 가속도 센서를 이용하여 실시간 보정 할 수 있는 방법을 제안한다. Blur현상은 획득한 이미지에서 발생하는 번짐 효과이다. 소형의 모바일 단말기는 사용자의 미세한 손 떨림에도 크게 흔들릴 수 있기 때문에 blur현상이 크게 나타나며, 이를 적절하게 보정할 수 있는 알고리즘이 필요하다. 본 논문에선 3축 가속도센서를 진자운동에 적용하여 출력결과의 신뢰성을 확보하였고, blur현상을 Uniform 분포와 Gaussian 분포로 모델링하였다. 실험을 통하여 실제 blur 현상이 Non-Gaussian 형태로 모델링됨을 확인하였고, 이 blur모델의 역과정인 deblurring 특성함수를 설계하였다. 이 특성함수에 3축 가속도센서에서 발생하는 미세한 떨림 정보를 적용하여 실험 이미지를 deblurring한 결과, 이미지 blur현상을 적절하게 보정할 수 있었다.
Keywords

SSE 명령어를 이용한 영상의 고속 전처리 알고리즘

Title SSE 명령어를 이용한 영상의 고속 전처리 알고리즘
Journal 대한전자공학회논문지 Vol. 49, No. 2 pp.65~77
Author 박은수, 최학남, 김학일
Abstract 본 논문에서는 SSE (Streaming SIMD Extensions) 명령어를 이용한 고속 영상처리 알고리즘을 제안한다. SSE 명령어를 지원하는 CPU는 128비트 크기의 XMM 레지스터를 보유하고 있으며 이에 속한 데이터는 SIMD(Single Instruction Multiple Data) 방식으로 한 번에 병렬로 처리 될 수 있다. 영상처리에서 폭넓게 활용되는 평균 필터, 소벨 수평방향 외곽선 검출, 이진침식 알고리즘을 SIMD 방식으로 효과적으로 처리 할 수 있는 알고리즘을 제시하였고, 수행 시간을 측정하였다. 보다 객관적인 수행 속도 평가를 위해 현재 많이 사용되고 있는 영상처리 라이브러리와의 수행 속도를 비교하였다. 비교에 사용된 라이브러리는 SISD(Single Instruction Single Data)방식으로 동작하는 OpenCV 1.0, SIMD 방식을 지원하는 고속 영상처리 라이브러리인 IPP 5.2와 MIL 8.0에서 각각 수행 시간을 측정하고 제안하는 알고리즘의 처리 속도와 비교하였다. 실험결과 제안하는 알고리즘은 SISD방식의 영상처리 라이브러리에 비해 평균 8배의 성능향상을 보였으며, SIMD 방식의 고속 영상처리 라이브러리와 비교 하였을 때 평균 1.4배의 성능향상을 보였다. 따라서 제안하는 알고리즘은 고가의 영상처리 라이브러리와 추가적인 하드웨어의 구입 없이도 고속으로 동작해야 하는 실제 영상 처리 어플리케이션에 효과적으로 적용될 수 있음을 보였다.
Keywords

CPU와 GPU의 병렬 처리를 이용한 고속 물체 인식 알고리즘 구현

Title CPU와 GPU의 병렬 처리를 이용한 고속 물체 인식 알고리즘 구현
Journal 제어, 로봇, 시스템학회 논문지 Vol. 15, No. 5 pp. 487-494
Author 박은수, 최학남, 김학일
Abstract This paper presents a fast feature extraction method for autonomous mobile robots utilizing parallel processing and based on OpenMP, SSE (Streaming SIMD Extension) and CUDA programming. In the first step on CPU version, the algorithms and codes are optimized and then implemented by parallel processing. The parallel algorithms are debugged to maintain the same level of performance and the process for extracting key points and obtaining dominant orientation with respect to key points is parallelized. After extraction, a parallel descriptor via SSE instructions is constructed. And the GPU version also implemented by parallel processing using CUDA based on the SIFT. The GPU-Parallel descriptor achieves an acceleration up to five times compared with the CPU-Parallel descriptor, but it shows the lower performance than CPU version. CPU version also speed-up the four and half times compared with the original SIFT while maintaining robust performance.
Keywords

Matching Performance-Based Comparative Study of Fingerprint Sample Quality Measures

Title Matching Performance-Based Comparative Study of Fingerprint Sample Quality Measures
Journal 정보보고학회 논문지 Vol.19, No,3 pp:11-25
Author  Hakil Kim
Abstract Fingerprint sample quality is one of major factors influencing the matching performance of fingerprint recognition systems. The error rates of fingerprint recognition systems can be decreased significantly by removing poor quality fingerprints. The purpose of this paper is to assess the effectiveness of individual sample quality measures on the performance of minutiae-based fingerprint recognition algorithms. Initially, the authors examined the various factors that influenced the matching performance of the minutiae-based fingerprint recognition algorithms. Then, the existing measures for fingerprint sample quality were studied and the more effective quality measures were selected and compared with two image quality software packages, (NFIQ from NIST, and QualityCheck from Aware Inc.) in terms of matching performance of a commercial fingerprint matcher (Verifinger 5.0 from Neurotechnologija). The experimental results over various Fingerprint Verification Competition (FVC) datasets show that even a single sample quality measure can enhance the matching performance effectively.
Keywords Fingerprint sample quality, Quality measure, Matching performance, Equal error rate

GPU를 이용한 Gabor Texture 특징점 기반의금속 패드 변색 분류 알고리즘

Title GPU를 이용한 Gabor Texture 특징점 기반의금속 패드 변색 분류 알고리즘
Journal 제어 ․ 로봇 ․ 시스템학회 논문지 제 15 권, 제 8 호 2009. 8 pp.778~785
Author 최 학 남, 박 은 수,  김 학 일
Abstract This paper presents a Gabor texture feature extraction method for classification of discolored Metal pad images using GPU(Graphics Processing Unit). The proposed algorithm extracts the texture information using Gabor filters and constructs a pattern map using the extracted information. Finally, the golden pad images are classified by utilizing the feature vectors which are extracted from the constructed pattern map. In order to evaluate the performance of the Gabor texture feature extraction algorithm based on GPU, a sequential processing and parallel processing using OpenMP in CPU of this algorithm were adopted.Also, the proposed algorithm was implemented by using Global memory and Shared memory in GPU. The experimental results were demonstrated that the method using Shared memory in GPU provides the best performance. For evaluating the effectiveness of extracted Gabor texture features, an experimental validation has been conducted on a database of 20 Metal pad images and the experiment has shown no mis-classification.
Keywords discolor, texture, GPU, P2C transform, pattern map

Plane-converging belief propagation을 이용한 고속 스테레오매칭

Title Plane-converging belief propagation을 이용한 고속 스테레오매칭
Journal 전자공학회 논문지 제48권 SP편 제2호, 2011. 03
Author  박은수, 김학일
Abstract Stereo matching is the research area that regarding the estimation of the distance between objects and camera using different view points and it still needs lot of improvements in aspects of speed and accuracy. This paper presents a fast stereo matching algorithm based on plane-converging belief propagation that uses message passing convergence in hierarchical belief propagation. Also, stereo matching technique is developed using GPU and it is available for real-time applications. The error rate of proposed Plane-converging Belief Propagation algorithm is similar to the conventional Hierarchical Belief Propagation algorithm, while speed-up factor reaches 2.7 times
Keywords Stereo matching, Belief propagation, Plane convergence, GPU

Robust iris segmentation via simple circular and linear filters

Title Robust iris segmentation via simple circular and linear filters
Journal Journal of Electronic Imaging (ISSN 1017-9900, IF:0.455) vol. 17, no. 4, 043027(Oct-Dec 2008)
SPIE and IS&T
Author Hakil Kim
Abstract This paper presents a robust method for iris segmentation. To detect the inner boundary of the iris, the method introduces a circular filter, which is specially designed to detect circular shapes in iris images. The inner boundary detection process consists of two steps. Firstly, the coarse center of the pupil is detected via the proposed filter. Then, a set of inner boundary points is found by detecting segment lines using the Radon transform in the rectangular areas which are converted from extracted arc regions with the coarse center. The inner boundary is finally determined by fitting the set of points to a circle using the least square method. In addition, the proposed method includes a process for extracting the outer boundary of the iris by applying a linear filter to the rectangular areas transformed from arc regions around the iris. Furthermore, a fast method for detecting eyelids after iris segmentation is presented in this paper. Experimental results over the CASIA iris databases show that the performance of the proposed methods is promising for iris images captured in unconstrained environments.
Keywords Iris segmentation, circular filter, circle detection, eyelid detection

 

Floor Segmentation by Computing Plane Normals from Image Motion Fields for Visual Navigation

Title Floor Segmentation by Computing Plane Normals from Image Motion Fields for Visual Navigation
Journal International Journal of Control, Automation, and Systems (2009) 7(5):788-798 – SCIE
Author Xue-Nan Cui, Hakil Kim
Abstract This paper proposes a method of detecting movable paths during visual navigation for a
robot operating in an unknown structured environment. The proposed approach detects and segments the floor by computing plane normals from motion fields in image sequences. A floor is a useful object for mobile robots in structured environments, because it presents traversable paths if existing static or dynamic objects are removed effectively. In spite of this advantage, it cannot be easily detected from 2D image. In this paper, some geometric features observed in the scene and assumptions about images are exploited so that a plane normal can be employed as an effective clue to separate the floor from the scene. In order to use the plane normal, two methods are proposed and integrated with a designed iterative refinement process. Then, the floor can be accurately detected even when mismatched point correspondences are obtained. The results of preliminary experiments on real data demonstrate the effectiveness of the proposed methods.
Keywords Floor detection, image motion estimation, layered image representation, mobile robot, visual navigation

 

High-resolution orientation field estimation based on multi-scale Gaussian filter

Title High-resolution orientation field estimation based on multi-scale Gaussian filter
Journal IEICE Electron. Express, Vol. 6, No. 24, pp.1781-1787, (2009)
Author  Hakil Kim
Abstract Orientation field plays the most important role in fingerprint recognition, Proposed in this paper is a novel approach of pixel-wise orientation field estimation using multi-scale, block-scale, and orientation-scale is developed for handling gradient vectors, coherence data, and orientation vectors, respectively. Experimental results on various FVC datasets show the proposed algorithm achieves accurate orientation field estimation which is robust to local defects, such as scar, low contrast, ridge discontinuity, smudged area, etc. with a low computational cost.
Keywords fingerprint recognition, Gaussian filter, orientation field