• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2006.05a
• /
• pp.401-404
• /
• 2006

본 논문에서는 영상의 중심이동과 독립성분분석에 의한 효율적인 표정 인식방법을 제안하였다. 여기서 중심이동은 얼굴영상의 1차 모멘트에 의한 전처리 과정으로 불필요한 배경을 배제시켜 계산시간의 감소 및 인식률을 개선하기 위함이다. 또한 독립성분분석은 얼굴표정의 특징으로 기저영상을 추출하는 것으로 고차의 통계성을 고려한 중복신호의 제거로 인식성능을 개선하기 위함이다. 제안된 방법을 320*243 픽셀의 48개(4명*6장*2그룹) 표정을 대상으로 Euclidean 분류척도를 이용하여 실험한 결과, 전처리를 수행치 않는 기존방법에 비해 우수한 인식성능이 있음을 확인하였다.

• Proceedings of the Korea Institutes of Information Security and Cryptology Conference
• /
• 1995.11a
• /
• pp.244-253
• /
• 1995

화상 내의 중복도를 의미하는 자기 상사성(self-similarity)을 제거하여 화상압출을 수행하는 프랙탈 화상압축 방식에 대한 연구가 최근에 활발히 진행되고 있다. 프랙탈 화상압축은 비가역 부호화 방식으로 화상을 일정블럭으로 분할하여 상사영역을 탐색하고 그 상사영역에 대한 파라메터를 압축을 위한 정보로 저장한다 전송하고자 하는 데이터가 화상과 기밀 데이터의 2가지가 혼재할 경우 일반적으로 각 데이터를 개별적으로 압축하고 전송가게 된다. 또한, 화상의 특징을 이용하여 기밀 데이터를 화상 내에 몰래 집어넣은 후, 그 화상을 압축하여 전송하는 방식이 있다. 본 논문에서는 이러한 방식과는 달리 기밀 데이터를 프랙탈 화상압축을 수행함과 동시에 예측부호화를 이용하여 제 3자가 인식할 수 없는 잡음의 형태로 합성하는 방법에 대하여 고찰하고 시뮬레이션 결과를 나타낸다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2004.11a
• /
• pp.583-586
• /
• 2004

본 논문에서는 자율학습의 속성을 가지는 주요성분분석을 이용한 얼굴인식 기법을 제안하였다. 이는 대용량의 입력 데이터를 통계적으로 독립인 특징들의 집합으로 변환시켜 중복신호를 제거하는 특성을 가지는 주요성분분석의 우수한 속성을 이용한 것이다. 제안된 기법을 Yale 얼굴영상 데이터베이스로부터 선택된 20개의 $320{\ast}243$ 픽셀의 영상을 대상으로 시뮬레이션한 결과, 주요성분의 개수에 따른 압축성능과 city-block, Euclidian, 그리고 negative angle(cosine)의 거리척도에 따른 인식에서의 분류성능에서 우수한 성능이 있음을 확인할 수 있었다.

• 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.

• Journal of KIISE:Software and Applications
• /
• v.37 no.2
• /
• pp.104-109
• /
• 2010

The lighting reflection is a common problem in image analysis and causes the many difficulties to extract distinct features in related fields. Furthermore, the problem grows in the rainy night. In this paper, we aim to remove light reflection effects and reconstruct a road surface without lighting reflections in order to extract distinct features. The proposed method utilizes a 3D analysis based on a multiple geometry using captured images, with which we can combine each reflected areas; that is, we can remove lighting reflection effects and reconstruct the surface. At first, the regions of lighting sources and reflected surfaces are extracted by local maxima based on vertically projected intensity-histograms. After that, a fundamental matrix and homography matrix among multiple images are calculated by corresponding points in each image. Finally, we combine each surface by selecting minimum value among multiple images and replace it on a target image. The proposed method can reduces lighting reflection effects and the property on the surface is not lost. While the experimental results with collected data shows plausible performance comparing to the speed, reflection-overlapping areas which can not be reconstructed remain in the result. In order to solve this problem, a new reflection model needs to be constructed.

• Journal of Broadcast Engineering
• /
• v.26 no.1
• /
• pp.26-38
• /
• 2021

In this paper, DLKF (Deep Learning Keypoint Filtering), the deep learning-based keypoint filtering method for the rapidization of the image registration method for remote sensing images is proposed. The complexity of the conventional feature-based image registration method arises during the feature matching step. To reduce this complexity, this paper proposes to filter only the keypoints detected in the artificial structure among the keypoints detected in the keypoint detector by ensuring that the feature matching is matched with the keypoints detected in the artificial structure of the image. For reducing the number of keypoints points as preserving essential keypoints, we preserve keypoints adjacent to the boundaries of the artificial structure, and use reduced images, and crop image patches overlapping to eliminate noise from the patch boundary as a result of the image segmentation method. the proposed method improves the speed and accuracy of registration. To verify the performance of DLKF, the speed and accuracy of the conventional keypoints extraction method were compared using the remote sensing image of KOMPSAT-3 satellite. Based on the SIFT-based registration method, which is commonly used in households, the SURF-based registration method, which improved the speed of the SIFT method, improved the speed by 2.6 times while reducing the number of keypoints by about 18%, but the accuracy decreased from 3.42 to 5.43. Became. However, when the proposed method, DLKF, was used, the number of keypoints was reduced by about 82%, improving the speed by about 20.5 times, while reducing the accuracy to 4.51.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.45 no.5
• /
• pp.38-47
• /
• 2008

The most commonly used location recognition system is the GPS-based approach. However, the GPS is inefficient for an indoor or urban area where high buildings shield the satellite signals. To overcome this problem, this paper propose the indoor positioning method using wavelet and neural network. The basic idea of proposed method is estimated the location using the received signal strength from wireless APs installed in the indoor environment. Because of the received signal strength of wireless radio signal is fluctuated by the environment factors, a feature that is strength of signal noise and error and express the time and frequency domain is need. Therefore, this paper is used the wavelet coefficient as the feature. And the neural network is used for estimate the location. The experiment results indicate 94.6% an location recognition rate.

• KIPS Transactions on Software and Data Engineering
• /
• v.3 no.12
• /
• pp.531-536
• /
• 2014

With the very large quantity of information available on the Internet, techniques for dealing with the abundance of documents have become increasingly necessary but the problem of processing information in the documents is still technically challenging and remains under study. Automatic document summary techniques have been considered as one of critical solutions for processing documents to retain the important points and to remove duplicated contents of the original documents. In this paper, we propose a document summarization technique that uses a fuzzy theory. Proposed summary technique solves the ambiguous problem of various features determining the importance of the sentence and the experiment result shows that the technique generates better results than other previous techniques.

• Proceedings of the Korea Institute of Convergence Signal Processing
• /
• 2006.06a
• /
• pp.29-32
• /
• 2006

본 논문에서는 기존의 생체인식에서 사용하지 않았던 방법으로 개인의 치열 영상을 이용하는 생체 인식 방법을 제안한다. 제안한 치열 인식 시스템은 데이터의 중복성 제거와 관측벡터의 차원 감소를 위하여 2D-DCT를 특징 파라미터로 사용하고, 음성인식 및 얼굴인식 분야에서 사용하는 EHMM 기술을 사용한다. EHMM은 3개의 super-state로 구성되며 각각의 super-state는 3개, 5개, 3개의 상태를 갖는 1D-HMM으로 구성된다. 치열인증 시스템의 성능 평가는 모델 훈련에 사용하지 않은 치열 영상으로 인식 실험하여 평가한다. 치열인식 실험에는 남자 10명과 여자 10명에 대하여 각각 10개의 이미지로 구성된 총 200개의 치열 영상을 사용한다. 치열인식 실험에서 제안한 치열인식 시스템의 인식률은 98.5%를 보였고, 참고문헌 [4]의 EHMM을 사용한 얼굴인식 시스템이 갖는 98%와 대등한 성능을 나타내는 것을 확인하였다.

• Journal of Digital Convergence
• /
• v.11 no.11
• /
• pp.343-349
• /
• 2013

Rough set theory comes to derive optimal rules through the effective selection of features from the redundancy of lots of information in data mining using the concept of equivalence relation and approximation space in rough set. The reduction of attributes is one of the most important parts in its applications of rough set. This paper purports to define a information-theoretic measure for determining the most important attribute within the association of attributes using rough entropy. The proposed method generates the effective reduct set and formulates the core of the attribute set through the elimination of the redundant attributes. Subsequently, the control rules are generated with a subset of feature which retain the accuracy of the original features through the reduction.