• International Journal of Contents
• /
• v.11 no.3
• /
• pp.47-53
• /
• 2015

Visual object tracking is a fundamental problem in the field of computer vision, as it needs a proper model to account for drastic appearance changes that are caused by shape, textural, and illumination variations. In this paper, we propose a feature-based visual-object-tracking method with a sparse representation. Generally, most appearance-based models use the gray-scale pixel values of the input image, but this might be insufficient for a description of the target object under a variety of conditions. To obtain the proper information regarding the target object, the following combination of features has been exploited as a corresponding representation: First, the features of the target templates are extracted by using the HOG (histogram of gradient) and LBPs (local binary patterns); secondly, a feature-based sparsity is attained by solving the minimization problems, whereby the target object is represented by the selection of the minimum reconstruction error. The strengths of both features are exploited to enhance the overall performance of the tracker; furthermore, the proposed method is integrated with the particle-filter framework and achieves a promising result in terms of challenging tracking videos.

• Journal of Internet Computing and Services
• /
• v.6 no.3
• /
• pp.159-170
• /
• 2005

OC-SVM(One Class Support Vector Machine) avoids solving a full density estimation problem, and instead focuses on a simpler task, estimating quantiles of a data distribution, i.e. its support. OC-SVM seeks to estimate regions where most of data resides and represents the regions as a function of the support vectors, Although OC-SVM is powerful method for data description, it is difficult to incorporate human subjective importance into its estimation process, In order to integrate the importance of each point into the OC-SVM process, we propose a fuzzy version of OC-SVM. In FOC-SVM (Fuzzy One-Class Support Vector Machine), we do not equally treat data points and instead weight data points according to the importance measure of the corresponding objects. That is, we scale the kernel feature vector according to the importance measure of the object so that a kernel feature vector of a less important object should contribute less to the detection process of OC-SVM. We demonstrate the performance of our algorithm on several synthesized data sets, Experimental results showed the promising results.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.36S no.1
• /
• pp.104-114
• /
• 1999

This paper proposes an efficient method for the automatic generation of personalized 3D face model from color image sequence. To detect a robust facial region in a complex background, moving color detection technique based on he facial color distribution has been suggested. Color distribution and edge position information in the detected face region are used to extract the exact 31 facial feature points of the facial description parameter(FDP) proposed by MPEG-4 SNHC(Synthetic-Natural Hybrid Coding) adhoc group. Extracted feature points are then applied to the corresponding vertex points of the 3D generic face model composed of 1038 triangular mesh points. The personalized 3D face model can be generated automatically in less then 2 seconds on Pentium PC.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.13 no.3
• /
• pp.266-274
• /
• 2003

This paper considers an incremental support vector learning for the abnormality detection problems. One of the most well-known support vector learning methods for abnormality detection is the so-called SVDD(support vector data description), which seeks the strategy of utilizing balls defined on the kernel feature space in order to distinguish a set of normal data from all other possible abnormal objects. The major concern of this paper is to modify the SVDD into the direction of utilizing the relation between the optimal solution and incrementally given training data. After a thorough review about the original SVDD method, this paper establishes an incremental method for finding the optimal solution based on certain observations on the Lagrange dual problems. The applicability of the presented incremental method is illustrated via a design example.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.7
• /
• pp.3756-3777
• /
• 2019

Recent internet development is helping malware researchers to generate malicious code variants through automated tools. Due to this reason, the number of malicious variants is increasing day by day. Consequently, the performance improvement in malware analysis is the critical requirement to stop the rapid expansion of malware. The existing research proved that the similarities among malware variants could be used for detection and family classification. In this paper, a Cross-Platform Malware Variant Classification System (CP-MVCS) proposed that converted malware binary into a grayscale image. Further, malicious features extracted from the grayscale image through Combined SIFT-GIST Malware (CSGM) description. Later, these features used to identify the relevant family of malware variant. CP-MVCS reduced computational time and improved classification accuracy by using CSGM feature description along machine learning classification. The experiment performed on four publically available datasets of Windows OS and Android OS. The experimental results showed that the computation time and malware classification accuracy of CP-MVCS was higher than traditional methods. The evaluation also showed that CP-MVCS was not only differentiated families of malware variants but also identified both malware and benign samples in mix fashion efficiently.

• Journal of Korea Multimedia Society
• /
• v.22 no.11
• /
• pp.1223-1232
• /
• 2019

The injection is the input method of the image feature vector from the encoder to the decoder. Since the image feature vector contains object details such as color and texture, it is essential to generate image captions. However, the bidirectional decoder model using the existing injection method only inputs the image feature vector in the first step, so image feature vectors of the backward sequence are vanishing. This problem makes it difficult to describe the context in detail. Therefore, in this paper, we propose the parallel injection method to improve the description performance of image captions. The proposed Injection method fuses all embeddings and image vectors to preserve the context. Also, We optimize our image caption model with Bidirectional Gated Recurrent Unit (Bi-GRU) to reduce the amount of computation of the decoder. To validate the proposed model, experiments were conducted with a certified image caption dataset, demonstrating excellence in comparison with the latest models using BLEU and METEOR scores. The proposed model improved the BLEU score up to 20.2 points and the METEOR score up to 3.65 points compared to the existing caption model.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.4 no.4
• /
• pp.491-508
• /
• 2010

In peer-to-peer (P2P) on-demand streaming networks, the alleviation of server load depends on reciprocal stream sharing among peers. In general, on-demand video services enable clients to watch videos from beginning to end. As long as clients are able to buffer the initial part of the video they are watching, on-demand service can provide access to the video to the next clients who request to watch it. Therefore, the key challenge is how to keep the initial part of a video in a peer's buffer for as long as possible, and thus maximize the availability of a video for stream relay. In addition, to address the issues of delivering data on lossy network and providing scalable quality of services for clients, the adoption of multiple description coding (MDC) has been proven as a feasible resolution by much research work. In this paper, we propose a novel caching scheme for P2P on-demand streaming, called Dynamic Buffering. The proposed Dynamic Buffering relies on the feature of MDC to gradually reduce the number of cached descriptions held in a client's buffers, once the buffer is full. Preserving as many initial parts of descriptions in the buffer as possible, instead of losing them all at one time, effectively extends peers’ service time. In addition, this study proposes a description distribution balancing scheme to further improve the use of resources. Simulation experiments show that Dynamic Buffering can make efficient use of cache space, reduce server bandwidth consumption, and increase the number of peers being served.

• Journal of Electrical Engineering and Technology
• /
• v.12 no.1
• /
• pp.363-372
• /
• 2017

This paper presents a novel face recognition algorithm based on the deep convolution neural network and key point detection jointed local binary pattern methodology to enhance the accuracy of face recognition. We firstly propose the modified face key feature point location detection method to enhance the traditional localization algorithm to better pre-process the original face images. We put forward the grey information and the color information with combination of a composite model of local information. Then, we optimize the multi-layer network structure deep learning algorithm using the Fisher criterion as reference to adjust the network structure more accurately. Furthermore, we modify the local binary pattern texture description operator and combine it with the neural network to overcome drawbacks that deep neural network could not learn to face image and the local characteristics. Simulation results demonstrate that the proposed algorithm obtains stronger robustness and feasibility compared with the other state-of-the-art algorithms. The proposed algorithm also provides the novel paradigm for the application of deep learning in the field of face recognition which sets the milestone for further research.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.10 no.6
• /
• pp.2730-2747
• /
• 2016

Due to the limitation of the bandwidth resource and capture resolution of depth cameras, low resolution depth maps should be up-sampled to high resolution so that they can correspond to their texture images. In this paper, a novel depth map up-sampling algorithm is proposed by exploiting the fractal internal self-referential feature. Fractal parameters which are extracted from a depth map, describe the internal self-referential feature of the depth map, do not introduce inherent scale and just retain the relational information of the depth map, i.e., fractal transforms provide a resolution-independent description for depth maps and could up-sample depth maps to an arbitrary high resolution. Then, an enhancement method is also proposed to further improve the performance of the up-sampled depth map. The experimental results demonstrate that better quality of synthesized views is achieved both on objective and subjective performance. Most important of all, arbitrary resolution depth maps can be obtained with the aid of the proposed scheme.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.3
• /
• pp.1595-1613
• /
• 2017

The IR camera and laser-based IR projector provide an effective solution for real-time collection of moving targets in RGB-D videos. Different from the traditional RGB videos, the captured depth videos are not affected by the illumination variation. In this paper, we propose a novel feature extraction framework to describe human activities based on the above optical video capturing method, namely spatial-temporal texture features for 3D human activity recognition. Spatial-temporal texture feature with depth information is insensitive to illumination and occlusions, and efficient for fine-motion description. The framework of our proposed algorithm begins with video acquisition based on laser projection, video preprocessing with visual background extraction and obtains spatial-temporal key images. Then, the texture features encoded from key images are used to generate discriminative features for human activity information. The experimental results based on the different databases and practical scenarios demonstrate the effectiveness of our proposed algorithm for the large-scale data sets.