Distributed compressed video sensing (DCVS) is a framework that integrates both compressed sensing and distributed video coding characteristics to achieve a low complexity video sampling. In DCVS schemes, motion estimation & motion compensation is employed at the decoder side, similarly to distributed video coding (DVC), for a low-complex encoder. However, since a simple BCS-SPL algorithm is applied to a residual arising from motion estimation and compensation in conventional MC-BCS-SPL (motion compensated block compressed sensing with smoothed projected Landweber) scheme, the reconstructed visual qualities are severly degraded in Wyner-Ziv (WZ) frames. Furthermore, the scheme takes lots of iteration to reconstruct WZ frames. In this paper, the conventional MC-BCS-SPL algorithm is improved to be operated in more effective way in WZ frames. That is, first, the proposed algorithm calculates a correlation coefficient between two reference key frames and, then, by selecting adaptively the reference frame, the residual reconstruction in pixel domain is performed to the conventional BCS-SPL scheme. Experimental results show that the proposed algorithm achieves significantly better visual qualities than conventional MC-BCS-SPL algorithm, while resulting in the significant reduction of the decoding time.

A fast cell contour extraction method using CUDA parallel processing technique is presented. The cell contour extraction is one of important processes to analyze cell information in pathology. However, conventional sequential contour extraction methods are slow for a huge high-resolution medical image, so they are not adequate to use in the field. We developed a parallel morphology operation algorithm to extract cell contour more quickly. The algorithm can create an inner contour and fail to extract the contour from the concave part of the cell. We solved these problems by subdividing the contour extraction process into four steps: morphology operation, labeling, positioning and contour extraction. Experimental results show that the proposed method is four times faster than the conventional one.

In order to recognize the facial expressions, good features that can express the facial expressions are essential. It is also essential to find the characteristic areas where facial expressions appear discriminatively. In this study, we propose a directional LBP feature for facial expression recognition and a method of finding directional LBP operation and feature region for facial expression classification. The proposed directional LBP features to characterize facial fine micro-patterns are defined by LBP operation factors (direction and size of operation mask) and feature regions through AdaBoost learning. The facial expression classifier is implemented as a SVM classifier based on learned discriminant region and directional LBP operation factors. In order to verify the validity of the proposed method, facial expression recognition performance was measured in terms of accuracy, sensitivity, and specificity. Experimental results show that the proposed directional LBP and its learning method are useful for facial expression recognition.

In this paper, we present a novel method to extract face representation based on multi-resolution spatial pyramid. In our method, a face is subdivided into increasingly finer sub-regions (local regions) and represented at multiple levels of histogram representations. To cope with misaligned problem, patch-based local descriptor extraction has been also developed in a novel way. To preserve multiple levels of detail in local characteristics and also encode holistic spatial configuration, histograms from all levels of spatial pyramid are integrated by using dimensionality reduction and feature combination, leading to our spatial-pyramid face feature representation. We incorporate our proposed face features into general face recognition pipeline and achieve state-of-the-art results on challenging face recognition problems.

Deep learning networks like Convolutional Neural Networks (CNNs) show successful performances in many computer vision applications such as image classification, object detection, and so on. For implementation of deep learning networks in embedded system with limited processing power and memory, deep learning network may need to be simplified. However, simplified deep learning network cannot learn every possible scene. One realistic strategy for embedded deep learning network is to construct a simplified deep learning network model optimized for the scene images of the installation place. Then, automatic training will be necessitated for commercialization. In this paper, as an intermediate step toward automatic training under fisheye camera environments, we study more precise human localization in fisheye images, and propose an accurate human localization method, Automatic Ground-Truth Labelling Method (AGTLM). AGTLM first localizes candidate human object bounding boxes by utilizing GoogLeNet-LSTM approach, and after reassurance process by GoogLeNet-based CNN network, finally refines them more correctly and precisely(tightly) by applying saliency object detection technique. The performance improvement of the proposed human localization method, AGTLM with respect to accuracy and tightness is shown through several experiments.

As most of the recent smart devices have NFC function the NFC mobile coupon will become one of the pervasive NFC applications. We need the secure NFC coupon protocols to issue and use NFC mobile coupon. In this paper, we analyze the security of the previous protocols and point out the problems of security. As the result of analysis, Premium M-coupon Protocol proposed by A. Alshehri and S. Schneider is the most secure but has unnecessary operations. We propose the Modified Premium M-coupon Protocol-1 with the unnecessary operations removed and show this protocol is secure by security analysis. Most of NFC mobile coupon protocols use the cryptography with the shared secret keys. We propose the Modified Premium M-coupon Protocol-2 without the shared secret keys and show this protocol is secure by security analysis.

With the development of the internet of things, sensor and device are can be applied to various scenario. Overall improving of the dependability index of internet of things is the ultimate goal. And reliability aims to increase the success rate of internet of things service delivery. Many studies about internet of things system have been made on the system to assess a dependability for providing reliable service to user, but it has difficult to reflect the user context for evaluating the device reliability. Also, most do not consider the availability of content information. In this paper, it proposed dependable device searching system in the internet of things environment. This system evaluates device dependability based on device status and measured data. Through the proposed system, it can be provided reliable context information for user-centric service.

The Pattern Sequence-based Forecasting (PSF) is an approach to forecast the behavior of time series based on similar pattern sequences. The innovation of PSF method is to convert the load time series into a label sequence by clustering technique in order to lighten computational burden. However, it brings about a new problem in determining the number of clusters and it is subject to insufficient similar days occasionally. In this paper we proposed an adaption of the PSF method, which introduces a new clustering index to determine the number of clusters and imposes a threshold to solve the problem caused by insufficient similar days. Our experiments showed that the proposed method reduced the mean absolute percentage error (MAPE) about 15%, compared to the PSF method.

Most of conventional DSS's(Digital Signage Systems) have been built based on LAMP framework. Recent researches have shown that MEAN or MERN stack framework is simpler, more flexible, faster and more suitable for web-based application than LAMP stack framework. In this paper, we propose a design and implementation of MEARN (ME(A+R)N) stack-based real-time digital signage system, MR-DSS, which supports handing real-time tasks like urgent/instant messaging, system status monitoring and so on, efficiently in addition to conventional digital signage CMS service tasks. MR-DSCMS, CMS of MR-DSS, is designed to provide most of its normal services by REST APIs and real-time services like urgent/instant messaging by Socket.IO base under MEARN stack environment. In addition to architecture description of components composing MR-DSS, design and implementation issues are clarified in more detail. Through experimental testing, it is shown that 1) MR-DSS works functionally well, 2) the networking load performance of MR-DSCMS's REST APIs is better compared to a well-known open source Xibo CMS, and 3) real-time messaging via Socket.IO works much faster than REST APIs.

Android's inter-application access enriches its application ecosystem. However, it exposes security vulnerabilities where end-user data can be exploited by attackers. While existing techniques have focused on minimizing the risks of inter-application access, they either suffer from inaccurate risk detection or are primarily available to expert users. This paper introduces a novel technique that automatically analyzes potential risks between a set of applications, aids end-users to effectively assess the identified risks by crowdsourcing assessments, and generates an access control policy which prevents unsafe inter-application access at runtime. Our evaluation demonstrated that our technique identifies potential risks between real-world applications with perfect accuracy, supports a scalable analysis on a large number of applications, and successfully aids end-users' risk assessments.

Humans have often been used gesture to communicate with each other. The communication between computer and person was also not different. To interact with a computer, we command with gesture, keyboard, mouse and extra devices. Especially, the gesture is very useful in many environments such as gaming and VR(Virtual Reality), which requires high specification and rendering time. In this paper, we propose a gesture recognition method based on CNN model to apply to gaming and real-time applications. Deep learning for gesture recognition is processed in a separated server and the preprocessing for data acquisition is done a client PC. The experimental results show that the proposed method is in accuracy higher than the conventional method in game environment.