• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.45 no.3
• /
• pp.123-134
• /
• 2008

In this paper, we propose a real-tin e face recognition system that is robust under various lighting conditions. Th Modified Census Transform algorithm that is insensitive to illumination variations is employed to extract local structure features. In a practical face recognition system, acquired images through a camera are likely to be blurred and some of them could be side face images, resulting that unacceptable performance could be obtained. To improve stability of a practical face recognition system, we propose a real-time algorithm that rejects unnecessary facial picture and makes use of recognition consistency between successive frames. Experimental results on the Yale database with large illumination variations show that the proposed approach is approximately 20% better than conventional appearance-based approaches. We also found that the proposed real-time method is more stable than existing methods that produces recognition result for each frame.

• Journal of Korea Multimedia Society
• /
• v.15 no.10
• /
• pp.1247-1256
• /
• 2012

There are many filter based image processing algorithms and they usually require a huge amount of computations and memory accesses making it hard to attain a real-time performance, expecially in embedded applications. In this paper, we propose a pipelined hardware structure of the filter based face detection algorithm to show that the real time performance can be achieved by hardware design. In our design, the whole computation is divided into three pipeline stages: resizing the image (Resize), Transforming the image (ICT), and finding candidate area (Find Candidate). Each stage is optimized by considering the parallelism of the computation to reduce the number of cycles and utilizing the line memory to minimize the memory accesses. The resulting hardware uses 507 KB internal SRAM and occupies 9,039 LUTs when synthesized and configured on Xilinx Virtex5LX330 FPGA. It can operate at maximum 165MHz clock, giving the performance of 108 frame/sec, while detecting up to 20 faces.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.16 no.10
• /
• pp.6708-6714
• /
• 2015

Various types of multimedia contents have been widely spread and distributed with the Internet that is easy to use. Meanwhile, Multimedia contents can bright a social problem because juveniles can access such harmful contents easily through the Internet. This paper proposes a method to determine if an input image is harmful or not, using an neural network. The proposed method first detects a face region from an input image through MCT features. The method then extracts skin color regions using color features and obtains candidate nipple areas from the extracted skin regions. Subsequently, we determine if the input image is harmful, by filtering out non-nipple regions using the artificial neural network. Experimental results show that the proposed method can effectively determine the harmfulness of input images.

• Journal of Digital Convergence
• /
• v.15 no.1
• /
• pp.11-22
• /
• 2017

The purpose of the study is to come up with counter plans to the spatial change caused by an aging society. To achieve the purpose of this study, research methods are conducted in the literature survey and the census data are compared. This study focuses on an aging society, the current status of space and related problems, based on political economic spatial concepts, and then presents five countermeasures as follows. First, the planning considering the aging populations. Second, increase in total fertility rate and increase population absorption. Third, increased economic vitality of the elderly due to increased participation in the production of senior citizens. Forth, establishment and implementation of regional development plan for the elderly. Fifth, needs to transform the spatial policies of the aged to prepare a large gap in space. The result of this paper proposes the need to change the living space policies and planning to avoid mismatching between them, reducing the aging speed simultaneously. The study is expected to contribute to the establishment of a space plan for areas where the aging population is rapidly increasing.

• KIPS Transactions on Software and Data Engineering
• /
• v.7 no.7
• /
• pp.267-274
• /
• 2018

This paper describes a head pose estimation method using PTZ(Pan-Tilt-Zoom) camera. When the external parameters of a camera is changed by rotation and translation, the estimated face pose for the same head also varies. In this paper, we propose a new method to estimate the head pose independently on varying the parameters of PTZ camera. The proposed method consists of 3 steps: face detection, feature extraction, and pose estimation. For each step, we respectively use MCT(Modified Census Transform) feature, the facial regression tree method, and the POSIT(Pose from Orthography and Scaling with ITeration) algorithm. The existing POSIT algorithm does not consider the rotation of a camera, but this paper improves the POSIT based on perspective projection in order to estimate the head pose robustly even when the external parameters of a camera are changed. Through experiments, we confirmed that RMSE(Root Mean Square Error) of the proposed method improve $0.6^{\circ}$ less then the conventional method.

• Journal of the Korea Society of Computer and Information
• /
• v.19 no.3
• /
• pp.25-35
• /
• 2014

This paper proposes a method of boosting-based classification for the purpose of real-time face detection. The proposed method uses depth images to ensure strong performance of face detection in response to changes in lighting and face size, and uses the depth difference feature to conduct learning and recognition through the I-MCTBoost classifier. I-MCTBoost performs recognition by connecting the strong classifiers that are constituted from weak classifiers. The learning process for the weak classifiers is as follows: first, depth difference features are generated, and eight of these features are combined to form the weak classifier, and each feature is expressed as a binary bit. Strong classifiers undergo learning through the process of repeatedly selecting a specified number of weak classifiers, and become capable of strong classification through a learning process in which the weight of the learning samples are renewed and learning data is added. This paper explains depth difference features and proposes a learning method for the weak classifiers and strong classifiers of I-MCTBoost. Lastly, the paper presents comparisons of the proposed classifiers and the classifiers using conventional MCT through qualitative and quantitative analyses to establish the feasibility and efficiency of the proposed classifiers.