Recently smart phones are replacing a number of existing mobile devices while gaining wide popularity. Taking pictures with smart phones became a big part of our daily lives as well as hobbies. However, smart phones have limited processing capabilities and display size compared to a PC. Therefore, it is hard to manage and explore photos in a single category basis when the number of photos in a phone increase. This paper provides an effective hierarchical photo exploring system. As generating a hierarchical model by extracting date/time and GPS data from smartphones, this system offers us with an efficient way to explore photos. This photo exploring system features (1) using user customizable virtual hierarchy (2) using hierarchical tree nodes merge (3) maximizing efficiency and convenience by using balanced hierarchy tree. It was designed and developed using a Google Android smart phone.
The number of released android software has been dramatically increasing over the last few years, thus users of android software have demanded to provide high quality and good services of the software. As a result, software testing methods for improving quality of software is growing important. Therefore, we propose a new software test case generation tool based on Record-Play Back (RPB) technique which is the most famous technique on GUI test techniques. The tool has a new methodology to resolve a problem in previous tools which spends lots of time to make a test case. Moreover, it also has a structure to automate some parts of the test case generation process. Those characteristics of the tool can help to reduce time and resources for generating a test case. Thus, the quality of software may be improved.
Aggregated search is a technique to search and assemble information from a variety of sources, within a single interface. This method has been adopted in most domestic and foreign portal sites such as Google, Yahoo, Naver and Daum. How to order the collections(news, blogs, images, etc.) in the aggregated search result is very important for finding relevant information quickly. In mobile environment, the influence of collection order is considerably enlarged due to small screen and limited interface. The collection order is usually decided based on click logs, but it is not effective for most long-tail keywords which have not enough click logs. In this paper, we propose a new framework to enhance the aggregated search result by referencing the click logs of semantically related keywords using an ontology. We apply query reduction methods into collection ordering to efficiently process multi-keyword queries. In addition, we exploit user profiles to improve the precision of the collection order. The experiments show that our framework enhances aggregated search results and reduces network traffic in mobile environment.
NTIS (National science and Technology Information Service) is a national R&D information portal service. NTIS gathers R&D information from the representative R&D management organization under each ministry. Due to the different methods adopted by each organization for the information connection, integration of such information through NTIS was inefficient. In this paper we compare information integration methods, and then design an information integration platform for the purpose of standardizing information integration. Based on the information integration platform, we also propose a model for NTIS information integration.
Various methods for efficient location based services in broadcast environments have been researched. These methods broadcast object data periodically and the clients receive the data and process a query with the data. Recently, continuous query processing methods considering the mobility of objects have been proposed. However, the client processes the query with past data because the object data broadcasts in a certain period and the data are decided on the beginning of the period while the objects move continuously. Therefore, the existing methods are not suitable to process a continuous query with the mobility of the objects. In this paper, we propose a new indexing method to provide a continuous query in broadcast environments. The proposed method uses the vector information of objects and estimates the location between the periods. It increases the accuracy by the estimation. In addition, we minimize the search area increased by the vectors according to the length of a period. To show the superiority of the proposed method, we evaluate its performance through various experiments.
Vehicular Delay Tolerant Networks (VDTNs) have been proposed to address the communication challenges through the store and forward techniques. These techniques employ the intermediate nodes, which take custody of the data and forward them when is possible, but this method could generate congestion and lost packets if the intermediate nodes are not properly selected. However, the selection of the intermediate nodes remains as a central challenge. Therefore, we propose an efficient destination-based algorithm for dynamic custodian management by selecting the intermediate nodes according to the importance of the message, destination, current location and speed of the vehicles. An experimental evaluation was performed with real traces of taxies in the Shanghai city to show the feasibility and assess the performance of the approach. The results revealed an outstanding performance in terms of delivery ratio.
With the advancement of statistical machine learning, various machine learning methods have been applied to dynamic analysis of multimodal streams. However, previous studies have limitations for tackling various real-world streams because they focus on utilizing very limited characteristics of certain domains such as repetition of fixed frames. In this paper, we introduce a generative model-based segmenting method in which a story segment of a video stream is estimated through the likelihood of a given model to explain incoming data without requiring prior knowledge. There exists a profound question of how to compare each segment's latent structure parameters. In the proposed model, this difficulty is circumvented by computing likelihood of a new frame given a story model. We apply the proposed method to distinguishing several story segments in a TV drama episode. We employ LDA (Latent Dirichlet Allocation) framework for generating a story segment model. The proposed method is validated by comparing its results with those of human estimation.
The edit-distance between two strings is the smallest number of operations required to transform one string into the other. The edit-distance problem for two languages is to find a pair of strings, each of which is from different language, with the minimum edit-distance. We consider the edit-distance problem for a regular language and a context-free language and present an efficient algorithm that finds an optimal alignment of two strings, each of which is from different language. Moreover, we design a faster algorithm for the edit-distance problem that only finds the minimum number of operations of the optimal alignment.
This paper deals with two critical issues in wireless sensor networks: reducing the end-to-end packet delivery delay and increasing the network lifetime through the use of cooperative communications. Here, we propose a delay- and energy-aware cooperative medium access control (DEC-MAC) protocol, which trades off between the packet delivery delay and a node's energy consumption while selecting a cooperative relay node. DEC-MAC attempts to balance the energy consumption of the sensor nodes by taking into account a node's residual energy as part of the relay selection metric, thus increasing the network's lifetime. The relay selection algorithm exploits the process of elimination and the complementary cumulative distribution function for determining the most optimal relay within the shortest time period. Our numerical analysis demonstrates that the DEC-MAC protocol is able to determine the optimal relay in no more than three mini slots. Our simulation results show that the DEC-MAC protocol improves the end-to-end packet delivery latency and the network lifetime significantly compared to the state-of-the-art protocols, LC-MAC and CoopMAC.
The energy consumption of a sensor network is significantly affected by task assignments to the nodes in the network. In this paper, a task assignment method to extend the lifetime of wireless sensor networks is presented that exploits task decomposition and transformation. The task assignment is formulated as an optimization problem by providing a cost function incorporating the task decomposition and transformation at the same time. To show the validity and feasibility of our proposed method, we implement a task assignment framework using a simulated annealing approach. The simulation results show that optimal assignments and task decomposition can significantly improve the lifetime of wireless sensor networks.
Mobile devices equipped with various sensors have the potential of providing context-aware services. Location is one of the most common forms of context, which can be applied to diverse applications. In this paper, we present methods for learning and predicting users’ routes between significant locations, e.g., home and workplaces, based on personal GPS data. A user’s significant locations and routes between them are learned by a set of rules as well as clustering. When the user is moving, our methods can predict which of the learned routes is being taken now. After the route prediction, the user’s next location can also be inferred. Our methods have been applied to the real GPS datasets from four subjects. For the next location prediction task, the achieved accuracy was 84.8%.
In this letter, we propose a multi-matrix programming model in GPU computation to deal with Hessenberg reduction problem. Hessenberg reduction problem is the most important step in eigenvalue problem, which is computationally expensive. Conventional method using GPU is inefficient in GPU resource usage when it deals with small matrix. The proposed method computes as many matrices as possible which maximally utilizes the GPU resources. Therefore, the proposed method achieves higher memory transfer rate between CPU and GPU, and massive parallel computation that is well fit for GPU computation. Experimental results show that the proposed outperforms the optimized commercial packages.
The performance of Graphic Processing Unit (GPU) has been improved dramatically. To improve the GPU performance continuously, a quantitative analysis on the various factors which degrade the GPU performance should be provided. In this paper, we divide the negative factors on the GPU performance into five types and analyze the impact of each factor quantitatively. According to our experimental results using GPGPU-SIM, memory overhead degrades the GPU performance by 12.6% when the mass-data applications are executed. In cases that computation-intensive applications are executed, the degradation of GPU performance caused by the interconnection overhead and register file overhead is measured as 3.5%.
The skyline of a multidimensional data set is defined to be the subset whose elements are optimal to the user's requirements, Thus, the skyline of a multidimensional data set is very useful for recommendation systems, Especially, in case the values of all the dimensional attributes of the set are pre-determined, skyline computation can be done in advance of the recommendation services, However, in a mobile environment where the user is moving around, his/her location should also be considered for recommendation services, In this study, in order to enhance the quality of the recommendation services, we define an extended skyline, propose an efficient scheme for computing the skyline, The effectiveness of the scheme is shown by experiments.
In order to effectively provide cloud computing, IT infrastructure which supports distributed file system and parallel data processing is essential. To this end, MapReduce framework has been widely used for distributed processing of large-scale data. MapReduce framework has been proven as an efficient way to construct distributed and parallel processing system at relatively low cost. However, it has the problem of single point of failure (SPOF) at JobTracker that is responsible for scheduling and assigning of all MapReduce tasks. When JobTracker has failed, the completion time of the MapReduce job is increased because the entire MapReduce tasks must be restarted. To resolve the above mentioned problem we designed and implemented JobTracker fault-tolerant mechanism for MapReduce framework. The performance of the mechanism is evaluated by using MapReduce testbed and fault-injection method. As a result, the average job completion time of the mechanism is dramatically reduced about 46.5%~64.4% compared to the result of a naive MapReduce.
Just-in-time compilation (JITC) and ahead-of-time compilation (AOTC) has been proposed to improve the performance of Java virtual machine (JVM). These techniques adopt Java specific optimizations as well as traditional compiler optimizations. One of Java specific optimizations is a null pointer check elimination, which is considered to be a mandatory optimization in most JVM, since it can achieve noticeable performance improvement by eliminating redundant overhead of checking null pointers. In this paper, we propose an extended null pointer check elimination using specialization. The proposed technique extends the scope of existing null pointer check elimination and can eliminate additional null pointer checks. In addition, the proposed technique can be adopted to existing Just-in-time compiler and Ahead-of-time compiler, because it preserves the semantic of existing null pointer check elimination optimization. We observed meaningful performance improvement with benchmark programs as well as real applications after applying the proposed optimization.
Many enterprises adopt virtual desktop technology to reduce total cost of ownership (TCO) according to the improvement of virtualization technologies that allow one physical machine to run multiple virtual machines. In addition, many cloud based virtual desktop services for common user have been introduced. To support this kind of virtual desktop, remote connection technique is essential. However, current remote connection technique needs a normal computation environment with keyboard, mouse and display. In this paper, we proposed a smartphone based connection broker between remote virtual desktop and local legacy devices. The broker can detect local devices that can be used as a resource of virtual machine such as LCD TV, keyboard and can connect them to remote virtual desktop.
As the share of embedded software in the development of weapon systems increases more and more, the reliability of weapon system is directly related to the software reliability. For delivering high reliable software to customers, LIG Nex1 has tried joint project aimed at improving the reliability of embedded software for years. According to policy changes about company-led-development of weapon systems. The share of outsourcing in the R&D project is increasing and it becomes essential to ensure the quality of co operation software. In this paper, we would like to introduce cases - securing the quality of embedded software, going on integrating weapon systems smoothly, and reducing maintenance costs by strengthening the partner's software engineering capabilities.
Due to the advances of wireless technologies, a variety of Radio Access Technologies (RATs) coexist. Under the heterogeneous RAT environment, mobile terminals such as smart phones, has are equipped with multiple wireless interfaces to support various RATs. However, the popularity of various wireless services causes lack of radio resources of wireless networks temporarily. This problem can be alleviated by common radio resource management (CRRM) for cooperating overlaid wireless networks. In this paper, we survey current standards for heterogeneous wireless networks to cooperate with each other and propose a novel CRRM architecture based on current standards.
The social network such as Facebook, Twitter has been rapidly popularized by the proliferation of mobile devices and their ubiquitous web connections. We present a novel way to share and synchronize media contents in the social network. The proposed system enables users to search and share their social media contents efficiently in mobile environment. Our system extracts metadata from social contents automatically during storing the social media contents. It also provides ranking techniques which are based on collaborative filtering to support personalized searches. Thereby, the proposed method helps users to share large scale and heterogeneous social media contents in social network.
