• Journal of KIISE:Software and Applications
• /
• v.30 no.3_4
• /
• pp.372-378
• /
• 2003

Limitations on the linear discriminant analysis (LDA) for face rerognition, such as the loss of generalization and the computational infeasibility, are addressed and illustrated for a small number of samples. The principal component analysis (PCA) followed by the LDA mapping may be an alternative that ran overcome these limitations. We also show that any schemes based on either mappings or template matching are vulnerable to image variations due to rotation, translation, facial expressions, or local illumination conditions. This entails the importance of a proper preprocessing that can compensate for such variations. A simple template matching, when combined with the geometrically correlated feature-based detection as a preprocessing, is shown to outperform mapping techniques in terms of both the accuracy and the robustness to image variations.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2018.05a
• /
• pp.484-486
• /
• 2018

Recently, with the increase in use of digital equipment in various fields, the importance of image and signal processing is increasing. However, many types of noise are generated during transmission and reception of digital signal, causing errors. For this reason, noise removal is mandatorily performed during pre-processing phase in many fields. In the present paper, noise is classified through noise evaluation, and noise removal is performed to remove impulse noise and noise with AWGN-added noise. And, proposed is an algorithm which utilizes modified Gaussian filter and directional effective pixels according to noise type. Simulation results show superior noise-removal characteristics, and for objective evaluation, compared with conventional methods.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2004.04a
• /
• pp.576-579
• /
• 2004

Eventhough the productivity has been improved remarkably by introducing automatic facilities, the 100％ inspection is necessary because the possibility to produce large amount of defective goods is also increased. Since it is extremely unreasonable that workers inspect very large amount of products as 100％ inspection, there has been many researches for the automatic inspection system. In this thesis, we develop an automatic detection system of suspension insulator's minutes cracks System The automatic detection system of suspension insulator's minute cracks： To detect the minute cracks of suspension insulators, images of the insulator are acquired with a progressive scan camera, rotating a suspension insulator on a turning table. And after the shadow and noises are eliminated by preprocessing techniques, we detect minute cracks using the features of them.

• Journal of the Korea Computer Graphics Society
• /
• v.8 no.1
• /
• pp.29-35
• /
• 2002

In virtual endoscopy, it is important to produce high quality perspective images in real-time. However, it is more significant to devise a navigation method that can make a virtual camera move through in human cavities such as colon and bronchus without collision and let the user control the camera intuitively. We propose an efficient navigation method, which generates 2D depth map during rendering the current frame, then determines position and direction of camera using the depth information. It offers collision-free navigation and allows us to control the camera as we want. Also it does not require preprocessing step and additional data structures.

• Proceedings of the KIEE Conference
• /
• 2015.07a
• /
• pp.1345-1346
• /
• 2015

본 논문에서는 보행자 및 배경 이미지로부터 HOG-PCA 특징을 추출하고 다항식 기반 RBFNNs(Radial Basis Function Neural Network) 패턴분류기과 최적화 알고리즘을 이용하여 보행자를 검출하는 시스템 설계를 제안한다. 입력 영상으로부터 보행자를 검출하기 위해 전처리 과정에서 HOG(Histogram of oriented gradient) 알고리즘을 통해 특징을 추출한다. 추출된 특징은 고차원이므로 패턴분류기 분류 시 많은 연산과 처리속도가 따른다. 이를 개선하고자 PCA (Principal Components Analysis)을 사용하여 저차원으로의 차원 축소한다. 본 논문에서 제안하는 분류기는 pRBFNNs 패턴분류기의 효율적인 학습을 위해 최적화 알고리즘인 PSO(Particle Swarm Optimization)을 사용하여 구조 및 파라미터를 최적화시켜 모델의 성능을 향상시킨다. 사용된 데이터로는 보행자 검출에 널리 사용되는 INRIA2005_person data set에서 보행자와 배경 영상을 각각 1200장을 학습 데이터, 검증 데이터로 구성하여 분류기를 설계하고 테스트 이미지를 설계된 최적의 분류기를 이용하여 보행자를 검출하고 검출률을 확인한다.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2012.10a
• /
• pp.894-897
• /
• 2012

We propose a effective method using the HOG (Histogram of Oriented Gradients) feature vector to track individual objects in an environment which multiple objects are moving. The proposed algorithm consists of pre-processing, object detection and object tracking. We experimented with six videos which have various trajectories and the movement. When occlusion between objects was occurred, we identified individual object by using center and predicted coordinates of moving objects. The algorithm shows 85.45% of tracking rate in the videos we experimented. We expect the proposed system is utilized in security systems which require the alalysis of the position and motion pattern of objects.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2012.10a
• /
• pp.75-77
• /
• 2012

In this paper, We proposed that real-time pupil detection using local binarization at each region of eyes in image. In image obtained a single low-resolution web-camera, we detect a region of face using haar-like feature and then detect each region of eyes depending upon the rate of width and height of region of face respectively. In each region of eyes, we detect the pupil after local preprocessing and binarizing. This pupil detection can be variously used for HCI(Human-Computer Interface) systems.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.32 no.10C
• /
• pp.1000-1009
• /
• 2007

This paper presents a content adaptive interpolation (CAI) for intra deinterlacing. The CAI consists of three steps: pre-processing, content classification, and adaptive interpolation. There are also three main interpolation methods in our proposed CAI, i.e. modified edge-based line averaging (M-ELA), gradient directed interpolation (GDI), and window matching method (WMM). Each proposed method shows different performances according to spatial local features. Therefore, we analyze the local region feature using the gradient detection and classify each missing pixel into four categories. And then, based on the classification result, a different do-interlacing algorithm is activated in order to obtain the best performance. Experimental results demonstrate that the CAI method performs better than previous techniques.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.15 no.8
• /
• pp.5263-5268
• /
• 2014

This paper suggests a fast human detection algorithm that can be applied to a high-resolution CCTV camera. Human detection algorithms, which used a HOG detector show high performance in the region of image processing. On the other hand, it is difficult to apply to real-time high resolution imaging because of its slow processing speed in the extracting figures of HOG. To resolve this problems, we suggest how to detect humans into two stages. First, candidates of a human region are found using background subtraction, and humans and non-humans are distinguished using a HOG detector only. This process increases the detection speed by approximately 2.5 times without any degradation in performance.

• Annual Conference of KIPS
• /
• 2017.11a
• /
• pp.781-784
• /
• 2017

Face recognition systems are characterized by low invasiveness of acquisition, and increasingly better reliability. Such systems may not be applied effectively, when the images are in low resolution (LR) as in the case that photos are taken from long distances, typically public surveillance. In theory, the high resolution (HR) image reconstructed from an LR face image, applying a super resolution (SR) method, can be used for face recognition. However, existing face SR algorithms may not give satisfactory results in face recognition. This article investigates the very low resolution face recognition problem and introduces a partial differential equation (PDE)-based SR method for a face recognition system of convolutional neural network (CNN).