Earticle

Nowadays, to take advantage of tags is a general tendency when users need to store or retrieve images on the Web. In this article, we introduce some approaches to calculate semantic importance of tags attached to Web images, and to make re-ranking the retrieved images according to them. We have compared the results from image re-ranking with two semantic providers, WordNet and Wikipedia. With the semantic importance of image tags calculated by using Wikipedia, we found the superiority of the method in precision and recall rate as experimental results.

This paper presents an advertising framework to ensure the targeted advertisement service for heterogeneous consumer devices with different capabilities and subscribers with different preferences by providing advertisement content selection and filtering mechanism in real time. To support scalable advertisement service in multi-screen service environment, we suggest a new advertising framework using XML schema based on ISO/IEC 21000-2 Digital Item Declaration Language.

This paper presents the LQ-Servo control with observer for TCP/AQM system considering an input time-delay. This controller structure is made by taking a traditional servo mechanism based on Linear Quadratic approach and by augmenting a new state variable to the feed forward loop. In order to estimate the state of TCP window packet size, the control structure is extended with observer based on Kalman filter.

This paper presents a more efficient method for medium access in a real-time ubiquitous sensor network. To reduce the latency for the delivery of event data from sensor nodes, a fixed-size contention window with a non-uniform probability distribution of transmitting in each slot is selected. Through a simulation using ns-2, a widely used network simulation package, it is shown that the proposed method can reduce the latency considerably as compared to conventional IEEE 802.11 MAC protocols for a sensor network with up to 256 nodes. It is also shown that the proposed MAC scheme realizes a latency similar to that realized by a decentralized CSMA-based MAC protocol for real-time ubiquitous sensor networks that are sensitive to latency.

This paper addresses a probabilistic approach to develop a robust face recognition system to partial variations such as occlusions. Based on the statistical feature extraction methods, we take the modular PCA method which nds eigenspace not for the set of whole images but for the sets of local image patches. Through the local feature extraction approach, we try to overcome the drawback of wholistic appearance-based conventional PCA, and consequently expect to improve robustness to local variations. The obtained local features are then applied to dene two probabilistic models for facial images: one for modeling distribution of features observed in usual facial images, and the other for modeling distribution of environmental variations observed in face image from one subject. The probabilistic model for general fa- cial images are used to evaluate the importance of each local patch. The probabilistic model for the environmental variations is used to evaluate the similarity between two local fea- tures. By combining two probabilistic models, we nally dene a distance measure between two face images, which can be applied for face recognition. Computational experiments on benchmark face database show that the proposed face recognition method can achieve re- markable robustness to local variations.

A database is an organized collection of data and the database data collection with DBMS is called a database system. A numerous database systems have been developed. In our previous work, we have introduced our design of database system for a VOD system. Nowadays computing power of mobile terminals is utterly astonishing and playing video on them became popular. So, this paper extends the previous work to mobile VOD and further discusses implementation of the design.

Biometrics has been used to verify intrusion and identification of intruder in video surveillance system. Biometrics was divided into traditional biometrics and soft biometrics. Traditional biometrics is a method using information on face, iris, and fingerprint. Soft biometrics is a method using information on gender, ethnicity, and height. Recently, many researchers have been researched human identification using soft biometrics in long distance environment. In this paper, we analyze a human identification with applying traditional biometrics or soft biometrics suitable for video surveillance system. Also, we propose a framework to solve problems such as lighting, occlusion, and shadow.

In this paper, a two step efficient pose based gender classification technique has been presented. In first step, the proposed technique uses different type of features for pose classification which classifies the input image as frontal or side pose image. Based on pose classification result, gender classification is performed in the second step. Two different types of gender classifiers have been proposed for both side pose and frontal images which use the image and pose classification result as input and classifies the image as male or female. The proposed technique is tested over standard datasets using well known quantitative measures. Results show that the proposed technique gives superior performance than the existing techniques.

In this paper, we take into considerations of the inverse problem for LQ-PID control. Solving the inverse problem, explicit formulas are developed such that the closed loop system is optimal by an LQ-PID controller. The inverse problem of LQ-PID control is analytically formulated via the frequency-domain and algebraic characterizations of LQR with respect to special forms of the state weighting matrix Q. We apply the LQ-PID control to AQM(Active Queue Management) routers from the viewpoint of the inverse problem. In the controller design procedure, the practical effectiveness of the proposed analysis is demonstrated by the numerical simulation of AQM routers.

To minimize the longitudinal strip thickness error in the hydraulic cold rolling industries and improve traditional AGC control scheme on this problem, we put forward a secure gauge control system based on Advanced RISC Machines (ARM) and intelligent PI control strategy. The traditional automatic gauge control (AGC) scheme depended on personal experience about the parameters. The new gauge control system decreases the longitudinal strip thickness error via fuzzy control and ARM-based computing. Besides, this new system can be smaller than the traditional AGC system, and so is much easier to transport. The Fuzzy control scheme can increase the robustness of the ARM-based gauge control system against some disturbances. The Matlab-based experimental simulations show that our secure gauge control system with Fuzzy control and ARM is better than the traditional one.

The electrocardiography (ECG) artifact corrupts the surface electromyography (sEMG) signals recorded from the trunk area We assessed the effectiveness of three methods used to remove the ECG. We compared the performance of bandpass filtering methods, the commonly used mathematical morphology operator (MMO) method and the Adaptive filter on both simulationed and real sEMG data. False positive and false negative errors increased with larger ECG to EMG ratios for all 3 methods. However, the Adaptive filter showed better performance as compared to bandpass filtering and the MMO particularly for low signal-to-artifact ratios of a lower computational load.