This paper presents a set of techniques used in a real-time visual surveillance system. The system is implemented on a low-cost embedded DSP platform that is designed to work with stationary video sources. It consists of detection, a tracking and a classification module. The detector uses a statistical method to establish the background model and extract the foreground pixels. These pixels are grouped into blobs which are classified into single person, people in a group and other objects by the dynamic periodicity analysis. The tracking module uses mean shift algorithm to locate the target position. The system aims to control the human density in the surveilled scene and detect what happens abnormally. The major advantage of this system is the real-time capability and it only requires a video stream without other additional sensors. We evaluate the system in the real application, for example monitoring the subway entrance and the building hall, and the results prove the system's superior performance.

Feature based stereo matching is an effective way to perform 3D building reconstruction. However, in urban scene, the cluttered background and various building structures may interfere with the performance of building reconstruction. In this paper, we propose a novel method to robustly reconstruct buildings on the basis of rectangle regions. Firstly, we propose a multi-scale linear feature detector to obtain the salient line segments on the object contours. Secondly, candidate rectangle regions are extracted from the salient line segments based on their local information. Thirdly, stereo matching is performed with the list of matching line segments, which are boundary edges of the corresponding rectangles from the left and right image. Experimental results demonstrate that the proposed method can achieve better accuracy on the reconstructed result than pixel-level stereo matching.

Multi-Instance Learning(MIL) performs well to deal with inherently ambiguity of images in multimedia retrieval. In this paper, an effective framework for Contented-Based Image Retrieval(CBIR) with MIL techniques is proposed, the effective mechanism is based on the image segmentation employing improved Mean Shift algorithm, and processes the segmentation results utilizing mathematical morphology, where the goal is to detect the semantic concepts contained in the query. Every sub-image detected is represented as a multiple features vector which is regarded as an instance. Each image is produced to a bag comprised of a flexible number of instances. And we apply a few number of MIL algorithms in this framework to perform the retrieval. Extensive experimental results illustrate the excellent performance in comparison with the existing methods of CBIR with MIL.

Fruition for a diverse uIT(Ubiquitous Information Technology) service is currently required through the convergence of what is known as ubiquitous computing technology. In recent years, with the rapid growth of the need for information services, heavier demands have been made upon the supply of high-quality multimedia services that require a high bandwidth. Today, various home network solutions have been proposed to manage these service smoothly and a system able to deal with various types of network traffic without delays. In addition, in established home network environments that mix wired and wireless networks, along with the steady evolution of wireless network technology and the earnest development of ubiquitous computing technology, sensor-based wireless networks have emerged as a core solution for home networks. In this paper, a sensor network-based home control system supporting the next generation home gateway system for ubiquitous home environments is designed and implemented.

Tree-based reliable multicast protocols provide scalability by distributing error-recovery tasks among several repair nodes. These repair nodes perform local error recovery for their receiver nodes using the data stored in their buffers. We propose a packet loss patterns adaptive feedback scheduling scheme to manage these buffers in an efficient manner. Under our scheme, receiver nodes send NAKs to repair nodes to request packet retransmissions only when the packet losses are independent events from other nodes. At dynamic and infrequent intervals, they also send ACKs to indicate which packets can be safely discarded from the repair node's buffer. Our scheme reduces delay in error recovery because the requested packets are almost always available in the repair node's buffers. It also reduces the repair node's workload because (a) each receiver node sends infrequent ACKs with non-fixed intervals and (b) their sending times are fairly distributed among all the receiver nodes.

In mobile ad-hoc wireless networks, one of the most important challenging issues is how to conserve energy, maximizing the lifetime of route(networks) in the view points of both power and mobility of nodes. However, many transmission methods presented in the previous works can not satisfy these two objectives simultaneously. To obtain these two goals, in this paper we propose an architecture to support power saving transmission services with route stability in mobile ad-hoc wireless networks. The proposed architecture consists of two parts, the underlying route stability method to support route(network) lifetime and the power saving transmission methods. The performance evaluation of the proposed architecture is achieved via simulation and analysis.

In this paper, we proposed an efficient short address allocation mechanism for WiBEEM devices using NAA(Next Address Available) algorithm. The proposed addressing mechanism is based upon the NAA information that is delivered over the beacons every time it is transmitted at the beginning of each super-frame. The NAA-based addressing mechanism is not a systematic way of allocating short addresses to newly joining devices and thus tree-routing cannot be supported. However, it has great advantages when U-City core services including U-Parking Lot System or ATIS(Advanced Traveler Information System) services that require high-speed mobility are considered. Moreover, the proposed addressing mechanism can provide significant expandability of the wireless network to various applications and fast device discovery.

In this paper, we present a virtual tourist system with realtime 3D visualization and the assistance of context information service. Our system enables a visitor to take a discovering tour on a virtual environment from a remote client by following navigator or by self-navigating. During the tour, the system provides immersive 3D graphics contents while supporting relevant information to the visitors corresponding to their positions in the virtual environment. When the visitors interact with interested objects, the context information service will also support introduction information for presenting about the objects. The introduction information based on text format is represented by a comfortable way-audio conversion to visitors in different languages depended on their preferences using TTS(Text-To-Speak) tool.

This paper proposes a new group mobility model for wireless communication. The mobility model considers the psychological and sociological behavior of each node and the perception of other nodes for describing interactions among a set of nodes. The model assumes no permanent membership of a group, capable of capturing natural behaviors as fork and join. It emulates a cooperative movement pattern observed in mobile ad hoc networks of military operation and campus, in which a set of mobile stations accomplish a cooperative motion affected by the individual behavior as well as a group behavior. The model also employs a physic model to avoid a sudden stopping and a sharping turning.