Earticle

Scent delivery systems or what are known as olfactory displays has been the most recent trend in human computer interaction, control and automation, virtual reality, augmented reality, e-commerce and many other fields in the computer world. These devices have the ability to generate smell combinations for some digital items in the computer virtual world. The idea is to deliver additional perception (olfactory perception) to the user to enrich his/her indulgence while interacting with the computer. The proposed olfactory display is similar to this concept but greatly differ in many aspects, using laser based atomization method to produce enough sudden heat to generate odors that cannot be generated using chemical materials these odors are smoke, gun powder, wood milling, cigarettes smoking, plastic melting and Incense. All olfactory displays researchers avoid involving these types of odors in their studies due to the difficulty of the generation and control of such odors and the absence of their correct and precise generation using chemical materials. The proposed system will provide a way of generating such odors while interactive game is being played by a user with minimum impact on the environment to avoid adaption to the user's olfactory system and polluting the environment that the user occupies.

In this paper, a Parallel Multiprocessor (PM) based on SIMT (Single Instruction and Multiple Threads) architecture is proposed. With co-issue architecture and multi-dimensional parallelism implemented in high-effective PM, Embedded Graphics Processing Unit (EGPU) provides great performance for various situations, such as general purpose computing, 3D scene rendering, and graphics processing. Application programs are departed into separated threads. Allocated by Thread Processing Unit (TPU), separated threads can be executed in parallel. Parallelism in different hierarchy and dimension are implemented by Multi-Dimensional Parallelism Processor (MD-PP), which has made a proper trade-off between performance and cost. Additionally, PM improves the hardware occupancy with its co-issue architecture and internal bus accessing mechanism to meet the demand of processing capability. Its unified shading architecture also helps to hide processing latency. PM can execute 4 basic operations in the best case and 2 in the worst case within each clock cycle. With 32nm process technology and 200MHz clock frequency, PM’s area is about 5104494um2, power consumption is about 101.838mW, and it can process nearly 28M vertices or fragments in average. Experimental results show that the MD-PP based PM can process data with high performance and get a balance between efficiency and hardware consumption simultaneously.

Given the importance of the textual information in content retrieval, it is desirable that the textual representation of educational videos contents in social media platforms like YouTube capture the semantics of what is really in content they represent. Such coherent textual representations are important in objective video content retrieval, repurposing, reuse and sense- making of the content. In this study,the Automatic Speech Recognition (ASR) in the video tracks was leveraged to supplement the insufficient video content representations done through video title alone. The Latent Dirichlet allocation (LDA) implementation of Gibb’s sampling topic modeling approach was used to evaluate the suitability of various textual representations for YouTube educational videos and extract the candidate topic that extends well the original YouTube keywords. The results show that in topics space, YouTube ASR script performs well as a representative textual source in dominant topic than the combined textual representations. The automatic keywords extension obtained using our method add value to applications that use tags for content discovery or retrieval

With the development of mobile terminal technology, the continuous range query of moving objects on road network has been widely applied in the field of transportation, military and communication. In practical applications, the sampling frequency of positioning equipment could not eliminate uncertainty, resulting in moving objects’ position uncertainty between two adjacent samples. The index existing for continues probabilistic range query are based on the centralized or the traditional cluster distributed environment. In this paper, we construct UPBI index structure for the continuous probability range queries on road network based on Hadoop firstly. Secondly, we design the continuous probability range query parallel algorithm considering moving objects’ position uncertainty on road network. Finally, we simultaneously give space constraint R-restrict and the probability calculation method. The experiment demonstrates that index and query algorithm proposed effectively solve the mass data problem about moving objects, and enhance query efficiency.

In order to increase the ability to track face movements with large head rotations, a 3D shape model is used in the system. In this paper, we present a robust nose tip detection method in 3D facial image that handling facial expression and hair occlusion. The 3D face smoothed by weighted median filter, the holes are filled by linear interpolation during the re-sampling phase and the 3D Gaussian filter is used to remove noise. Since the database, mesh contains unimportant parts like neck, shoulder, clothes and hair that can also change the overall appearance of a face. We propose a 3D mask to cut the face and crop useful part, which helps us to achieve sufficient accuracy for noise detection. Nose localization is one of the most significant tasks of any facial classification system, compared to other facial landmarks. We proceed to detect the point N from the front image following the assumption that the relevant point has the highest value of the Z axis. In this framework, the face model is determined from a frontal 3D face image. In this experiment, the performance rate is improved from 90.1% to 98.3%. As indicated by the experimental result, the proposed binary mask with maximum intensity method provides a significant improvement in performance of nose tip detection, also it success in different facial expression.

To further enhance the distribution uniformity and extensiveness of the solution sets and to ensure effective convergence of the solution sets to the Pareto front, we proposed a MOEA approach based on a clustering mechanism. We named this approach improved multi-objective evolutionary algorithm (LMOEA). This algorithm uses a clustering technology to compute and maintain the distribution and diversity of the solution sets. A fuzzy C-means clustering algorithm is used for clustering individuals. Finally, the LMOEA is applied to solve the classical multi-objective knapsack problems. The algorithm performance was evaluated using convergence and diversity indicators. The proposed algorithm achieved significant improvements in terms of algorithm convergence and population diversity compared with the classical NSGA-II and the MOEA/D.

Model Driven Development (MDD) and Domain Specific Modeling (DSM) have been widely used in information system domains and achieved success in many open or in-house scenarios. But its application in the game domain is seldom and immature. In our research, we identified three issues that should be considered carefully in order to play the strength of MDD in the game development environment to a larger extend: 1) structured domain analysis should be done to assure the size and familiarity of the domain; 2) adapted process should be designed to save cost and support evolution; and 3) proper tools (especially language workbenches) should be evaluated and utilized to ease DSM tasks and accelerate iterations. In this paper, we explain these three issues and illustrate our solutions to them by presenting the development details (both technical and procedural) of one pervasive game case. We evaluate the gains and costs by involving MDD into the game development process. We reflect on the issues we have met, and discuss possible future works as well.

In this paper, in order to improve scalability of conferencing system in SIP(session initiation protocol) environment, a distributed large-scale conferencing system architecture with dynamic server allocation has been proposed. For this architecture, a new extended conference information data format has been presented to distribute system load dynamically to multiple conference servers. In order to control system load and conference object, the proposed extended conference information data format has been designed to include new added elements and attributes. The extended conference event package has been used to synchronize overall system conference information data between conference servers. This paper also shows detailed exchange procedures of SIP messages between conference servers and participants for distributing system load dynamically. An extensive experimental measurements have been done to verify the performance of the proposed distributed conferencing system architecture. The results show that the proposed architecture improves highly scalability and flexibility of the conferencing system

With the fast development of image processing technique, segmentation related issues have gained special attention in the research community. In this paper, an improved FCM combining mean shift algorithm is proposed to improve the segmentation visual effects and efficiency of traditional FCM. Initially, image segmentation into many small homogeneous area using the mean shift algorithm is conducted, segmentation and uniform area, rather than pixels as the new node. Then, image local entropy is adopted to describe the new nodes spatial and gray feature. Finally, an exponential function which is able to well simulate human nonlinear visual reaction was used to measure the similarity between the new node and the cluster center node. The experimental result shows the effectiveness and robustness of our proposed FCM, further potential research is also discussed.