Showing papers in "IEEE Transactions on Consumer Electronics in 2012"

Adaptive streaming of audiovisual content using MPEG DASH

Abstract: HTTP streaming has become a cost effective means for multimedia delivery nowadays. In this paper, we study the use of MPEG DASH standard to stream audiovisual content. We first describe a novel estimation method for connection throughput. Then, employing the extensibility feature of DASH syntax, we present a systematic method for selecting the best audio and video alternatives given the estimated throughput. The experiments show that our solution is stable to short-term fluctuations while responding quickly to large fluctuations.

Fast multi-exposure image fusion with median filter and recursive filter

Abstract: This paper proposes a weighted sum based multi-exposure image fusion method which consists of two main steps: three image features composed of local contrast, brightness and color dissimilarity are first measured to estimate the weight maps refined by recursive filtering. Then, the fused image is constructed by weighted sum of source images. The main advantage of the proposed method lies in a recursive filter based weight map refinement step which is able to obtain accurate weight maps for image fusion. Another advantage is that a novel histogram equalization and median filter based motion detection method is proposed for fusing multi-exposure images in dynamic scenes which contain motion objects. Furthermore, the proposed method is quite fast and thus can be directly used for most consumer cameras. Experimental results demonstrate the superiority of the proposed method in terms of subjective and objective evaluation.

An intelligent self-adjusting sensor for smart home services based on ZigBee communications

Abstract: Wireless sensor networks (WSNs) have been becoming increasingly essential in recent years because of their ability to manage real-time situational information for various novel services. Recently, the scope of WSN technologies has been expanded to places such as the home, in order to provide the residents with various intelligent services, such as home automation services or home energy management services. However, due to their architectural constraints, such as the trade-off between the performance and cost, WSNs are not effectively implemented in home environments. Therefore, this paper proposes a ZigBee-based intelligent self-adjusting sensor (ZiSAS) in order to address these concerns. This paper presents a situation-based selfadjusting scheme, an event-based self-adjusting sensor network, and hardware and middleware implementation. We also introduce some smart home services using the proposed system. We implemented our system in real test bed and conducted an experiment. Our experiment shows that we reduce the system's energy consumption.

Depth video-based human activity recognition system using translation and scaling invariant features for life logging at smart home

Abstract: Video-based human activity recognition systems have potential contributions to various applications such as smart homes and healthcare services. In this work, we present a novel depth video-based translation and scaling invariant human activity recognition (HAR) system utilizing R transformation of depth silhouettes. To perform HAR in indoor settings, an invariant HAR method is critical to freely perform activities anywhere in a camera view without translation and scaling problems of human body silhouettes. We obtain such invariant features via R transformation on depth silhouettes. Furthermore, in R transforming depth silhouettes, shape information of human body reflected in depth values is encoded into the features. In R transformation, 2D feature maps are computed first through Radon transform of each depth silhouette followed by computing 1D feature profile through R transform to get the translation and scaling invariant features. Then, we apply Principle Component Analysis (PCA) for dimension reduction and Linear Discriminant Analysis (LDA) to make the features more prominent, compact and robust. Finally, Hidden Markov Models (HMMs) are used to train and recognize different human activities. Our proposed system shows superior recognition rate over the conventional approaches, reaching up to the mean recognition rate of 93.16% for six typical human activities whereas the conventional PC and IC-based depth silhouettes achieved only 74.83% and 86.33% ,while binary silhouettes-based R transformation approach achieved 67.08% respectively. Our experimental results show that the proposed method is robust, reliable, and efficient in recognizing the daily human activities.

Real-time eye gaze tracking for gaming design and consumer electronics systems

Abstract: Real time face detection combined with eyegaze tracking can provide a means of user input into a gaming environment. Game and CE system designers can use facial and eye-gaze information in various ways to enhance UI design providing smarter modes of gameplay interaction and UI modalities that are sensitive to a users behaviors and mood. Here we review earlier approaches, using wearable sensors, or enhanced IR illumination. Our technique only requires video feed from a low-resolution user-facing camera. The algorithm is described and some comparative results on a range of embedded hardware are provided. The potential for using eye-gaze as a means of direct user input and improving the accuracy of estimation accordingly is also discussed.

Optimal design of a wireless power transfer system with multiple self-resonators for an LED TV

Abstract: This paper proposes an optimal design method in an asymmetric wireless power transfer (WPT) system for a 150 watt LED TV. The WPT system has three self-resonators: a Tx resonator, an Rx resonator, and an intermediate resonator. The Tx and Rx resonators are perpendicular and offset, respectively, to the intermediate resonator in the geometry. For optimal design, the WPT system is analyzed using an equivalent circuit. In particular, a calculation method for mutual inductance in the system is expressed. The calculation results of mutual inductance are used to determine the optimal position of each self-resonator for maximizing the power transfer efficiency. For verification, a WPT system for a 150 watt, 47 inch LED TV is fabricated at 250 kHz. The WPT system exhibits wireless power transfer efficiency of 80%.

Intelligent cloud home energy management system using household appliance priority based scheduling based on prediction of renewable energy capability

Abstract: Recent advances in micro-grid and distributed renewable energy have facilitated more efficient home energy management systems. However, due to characteristics of renewable energy such as intermittent energy generation, home energy management systems are inefficient and the recent systems therefore cannot be successfully applied to existing home. Therefore, this paper proposes intelligent cloud home energy management system (iCHEMS), considering these issues. iCHEMS assigns dynamic priority to a household appliance according to the type of appliance and its current status. In accordance with the assigned priority, the use of household appliances is scheduled considering renewable energy capability. We implemented iCHEMS in the test bed and conducted an experiment to verify the efficiency of the proposed system. The results show that the proposed system reduces the average total power consumption by up to 7.3 percent.

Employing a RGB-D sensor for real-time tracking of humans across multiple re-entries in a smart environment

Abstract: The term smart environment refers to physical spaces equipped with sensors feeding into adaptive algorithms that enable the environment to become sensitive and responsive to the presence and needs of its occupants. People with special needs, such as the elderly or disabled people, stand to benefit most from such environments as they offer sophisticated assistive functionalities supporting independent living and improved safety. In a smart environment, the key issue is to sense the location and identity of its users. In this paper, we intend to tackle the problems of detecting and tracking humans in a realistic home environment by exploiting the complementary nature of (synchronized) color and depth images produced by a low-cost consumer-level RGB-D camera. Our system selectively feeds the complementary data emanating from the two vision sensors to different algorithmic modules which together implement three sequential components: (1) object labeling based on depth data clustering, (2) human re-entry identification based on comparing visual signatures extracted from the color (RGB) information, and (3) human tracking based on the fusion of both depth and RGB data. Experimental results show that this division of labor improves the system?s efficiency and classification performance.

An embedded real-time finger-vein recognition system for mobile devices

Abstract: With the development of consumer electronics, the demand for simple, convenient, and high-security authentication systems for protecting private information stored in mobile devices has steadily increased In consideration of emerging requirements for information protection, biometrics, which uses human physiological or behavioral features for personal identification, has been extensively studied as a solution to security issues However, most existing biometric systems have high complexity in time or space or both, and are thus not suitable for mobile devices In this paper, we propose a real-time embedded finger-vein recognition system for authentication on mobile devices The system is implemented on a DSP platform and equipped with a novel finger-vein recognition algorithm The proposed system takes only about 08 seconds to verify one input finger-vein sample and achieves an equal error rate (EER) of 007% on a database of 100 subjects The experimental results demonstrate that the proposed finger-vein recognition system is qualified for authentication on mobile devices

Enhancing privacy and dynamic federation in IdM for consumer cloud computing

Abstract: Consumer cloud computing paradigm has emerged as the natural evolution and integration of advances in several areas including distributed computing, service oriented architecture and consumer electronics. In this complex ecosystem, security and identity management challenges have cropped up, given their dynamism and heterogeneity. As a direct consequence, dynamic federated identity management with privacy improvements has arisen as an indispensable mechanism to enable the global scalability and usability that are required for the successful implantation of Cloud technologies. With these requirements in mind, we present an IdM architecture based on privacy and reputation extensions compliance with the SAMLv2/ID-FF standards1.

Building lighting automation through the integration of DALI with wireless sensor networks

Abstract: This paper focuses on the integration of Digital Addressable Lighting Interface (DALI) devices in wireless sensor networks. Since different manufacturers usually deal with one aspect of building automation - e.g. heating ventilation and air conditioning, lighting control, different kinds of alarms, etc. - final building automation system has different subsystems which are finally taken to an integrated building management system. The cost of this process is consequently increased due to additional hardware investment. Our main purpose is to provide the end consumer with an economical fully centralized system in which home appliances are managed by an IEEE 802.15.4-based wireless sensor network. Not only is it necessary to focus on the initial investment, but maintenance and energy consumption costs must also be considered. This paper explains the developed system along with a brief introduction to usual building automation protocols. Finally it presents future work in this field.

HEVStream: a framework for streaming and evaluation of high efficiency video coding (HEVC) content in loss-prone networks

Abstract: High Efficiency Video Coding (HEVC) is the next generation video compression standard currently under development within the ITU-T/ISO sponsored Joint Collaborative Team on Video Coding (JCT-VC). The standardization, and eventual adoption, of HEVC will contribute significantly to the future development of many consumer devices. Areas such as broadcast television, multimedia streaming, mobile communications and multimedia/video content storage will all be impacted by implementation of the emerging HEVC standard. Up to this point in time the research focus of HEVC has been on improvements to video compression efficiency and little work has been conducted into streaming of HEVC. In this work we consider the practical barriers to HEVC streaming in realistic environments and propose HEVStream, a streaming and evaluation framework for HEVC encoded content. Our framework fills the current gap in enabling networked HEVC visual applications and permits the implementation, testing and evaluation of HEVC encoded video streaming under a range of packet loss, bandwidth restriction and network delay scenarios in a realistic testbed environment. We provide a basic error concealment method for HEVC to overcome limitations within the decoder and an RTP packetisation format for HEVC Network Abstraction Layer (NAL) units. Comprehensive results of HEVC streaming experiments under various network circumstances are reported. These results provide an insight into the reduction in picture quality, measured as peak signal to noise ratio (PSNR), that can be expected under a wide range of network constraint and packet loss conditions. We report an average loss of 3.61dB when a bandwidth reduction of 10% is applied. We believe that this work will be amongst the first to report on successful design and implementation of HEVC network applications, and evaluation of the effects of network constraints or limitations on the quality of HEVC encoded video streams.

Two hybrid positioning system design techniques with lighting LEDs and ad-hoc wireless network

Abstract: Two design techniques for more accurate and more convenient hybrid positioning system with visible light communication (VLC) and ad-hoc wireless network infrastructure are proposed, in order to overcome the problems of high estimation error, high cost, and limited service range of the conventional positioning techniques. First method is based on a non-carrier VLC based hybrid positioning technique for applications involving of low data rate optical sensing and narrow-range visible light reception from transmitter, and long-range positioning. The second method uses a 4 MHz carrier VLC-based hybrid positioning technique for a high data rate optical sensing and wide-range visible light receiving from transmitter, and midrange positioning applications. In indoor environments with obstacles where there are longrange 77.314m and mid-range 23.68m distances between an observer and a target respectively, the hybrid positioning systems developed with two design techniques are tested, and measured results are analyzed and presented in this paper.

Fast feature-based video stabilization without accumulative global motion estimation

Abstract: This paper presents a novel digital video stabilization approach that provides both efficiency and robustness. In this approach, features of each frame are first detected by the oriented features from accelerated segment test (FAST) method, and then the corresponding features between consecutive frames are matched by a very fast binary descriptor which is based on the rotated binary robust independent elementary features (BRIEF). The oriented FAST combined with the rotated BRIEF, which is called ORB, is very efficient in feature detection and matching, and can be used to speed up the motion estimation without any hardware acceleration. In addition, an improved motion smoothing method is proposed to smooth affine model based motion parameters without accumulative global motion estimation. Unlike the conventional method, the proposed method uses unstable input frames and stabilized output frames instead of original input frames to estimate motion parameters directly, allowing for more desirable motion parameters. Experiments with a variety of videos demonstrate that the proposed approach is both efficient and robust.

Real-time household load priority scheduling algorithm based on prediction of renewable source availability

Abstract: We propose a real-time household load priority scheduling algorithm based on renewable source availability prediction to maximize the benefits of renewable sources and minimize the total cost of energy consumption with consumers' comfort constraints. Home appliances are assigned dynamic priority according to their different energy consumption modes and their corresponding status. Hour-byhour weather forecast is considered to predict the availability of the renewable sources. Based on the allocated priority, home appliances are scheduled according to the predicted output of renewable sources and the forecast electricity market price. In addition, to effectively schedule appliances according to the real-time output of renewable sources and the electricity market price changes, which generally deviate from the corresponding forecasting, an algorithm for realtime household load scheduling is proposed and its benefits on cost and energy efficiency are discussed as well.

Adaptive virtual environments for neuropsychological assessment in serious games

Abstract: While advances in military relevant simulations provide potential for increasing assessment of Soldier readiness to Return-to-Duty (eg, following a blast injury), little has been done to develop these simulations into adaptive virtual environments (AVE) for improved neurocognitive and psychophysiological assessment Adaptive assessments offer the potential for dynamically adapting the difficulty level specific to the patient's knowledge or ability We present an iteration of the Virtual Reality for Cognitive Performance and Adaptive Treatment (VRCPAT) that proffers a framework for adapting scenarios in a game engine based upon the user's neurocognitive (task performance) and psychophysiological (eg, heart rate, skin response, heart rate, and pupil diameter) states

Quality metric to assess video streaming service over TCP considering temporal location of pauses

Abstract: There is a wide range of video services over complex transmission networks, and in some cases end users fail to receive an acceptable quality level. In this paper, the different factors that degrade users' quality of experience (QoE) in video streaming service that use TCP as transmission protocol are studied. In this specific service, impairment factors are: number of pauses, their duration and temporal location. In order to measure the effect that each temporal segment has in the overall video quality, subjective tests. Because current subjective test methodologies are not adequate to assess video streaming over TCP, some recommendations are provided here. At the application layer, a customized player is used to evaluate the behavior of player buffer, and consequently, the end user QoE. Video subjective test results demonstrate that there is a close correlation between application parameters and subjective scores. Based on this fact, a new metrics named VsQM is defined, which considers the importance of temporal location of pauses to assess the user QoE of video streaming service. A useful application scenario is also presented, in which the metrics proposed herein is used to improve video services.

Design and implementation of a fall monitor system by using a 3-axis accelerometer in a smart phone

Abstract: In this paper we use the accelerometer of a smart phone to design and implement a fall monitor with GPS function for the user. We not only analyze the change of acceleration but also analyze the six typical actions of humans. These are going upstairs, going downstairs, standing up, sitting down, running and jumping. Then we compare the six actions with the characteristics of a fall. These are weightlessness, impact, immobility and overturning of the body. Our design is based on an open source system platform and the accelerometer in the smart phone. If there has been a fall motion our design provides the GPS functionality to obtain the location of the user and draws the help path on both the smart phones and the computers of his family, friends or help center immediately.

An efficient multiple object detection and tracking framework for automatic counting and video surveillance applications

Abstract: Automatic visual object counting and video surveillance have important applications for home and business environments, such as security and management of access points. However, in order to obtain a satisfactory performance these technologies need professional and expensive hardware, complex installations and setups, and the supervision of qualified workers. In this paper, an efficient visual detection and tracking framework is proposed for the tasks of object counting and surveillance, which meets the requirements of the consumer electronics: off-the-shelf equipment, easy installation and configuration, and unsupervised working conditions. This is accomplished by a novel Bayesian tracking model that can manage multimodal distributions without explicitly computing the association between tracked objects and detections. In addition, it is robust to erroneous, distorted and missing detections. The proposed algorithm is compared with a recent work, also focused on consumer electronics, proving its superior performance.

Robust exemplar-based inpainting algorithm using region segmentation

Abstract: This paper proposes a robust exemplar-based image inpainting algorithm using region segmentation. Exemplar-based inpainting methods iteratively search the source region and fill the missing or damaged region, i.e., target region, with the most similar patch in the source region. The proposed method uses segmentation map to improve the performance of robust inpainting, in which a segmentation method is used to utilize spatial information in the source region. With the segmentation map, the proposed method automatically selects parameter values of the robust priority function, adaptively determines patch size, and reduces search region. Experimental results with a number of test images show the effectiveness of the proposed method.

No-reference video quality assessment in the compressed domain

Abstract: In this paper, a novel no-reference video quality assessment algorithm in the compressed domain is introduced. The proposed algorithm takes into account of three key factors; the quantization parameter, the motion, and the bit allocation factor which are calculated using the information extracted from the compressed bitstream. Characteristics of human visual system are also taken into consideration. In contrast to conventional video quality assessment metrics which measure the video quality frame by frame, the video sequence is processed by this algorithm as a whole. The proposed algorithm has been tested on the H.264/AVC bitstreams in LIVE Video Quality Database. Furthermore, bitstreams encoded by JM 17.0 with different configurations are also tested to examine the generality. The experiment results show that the proposed algorithm has better performance than many conventional methods in terms of both the prediction accuracy and the computation complexity.

Verity: an ambient assisted living platform

Abstract: A new and innovative system intended for the monitoring of the elderly and those requiring passive care is developed, with a wrist-wearable wireless sensor node and a mobile platform to form a remotely accessible body area network. The system utilises a state-of-the art SoC (System on Chip), using an ultra-low power sensor interface and RF communication. Ambient/skin temperatures, accelerations and heart rate of the wearer are measured for real-time monitoring. Due to the high dimension and nonlinearity of the collected sensor data, a classifier with a dimension reduction technique is proposed for robust behaviour identification of a wearer. Experiments verify that the proposed classifier outperforms the conventional classifiers in its one-pass training and with higher distinguishing capability. The device functions unobtrusively, whilst giving the user peace of mind in the knowledge that their safety is being monitored.

Robust people counting system based on sensor fusion

Abstract: This paper presents a novel robust people counting system based on fusing the depth and vision data. Conventional algorithms utilize the monoscopic or stereoscopic vision data to count people. However, these vision-based people counting methods often fail due to occasional illumination change and crowded environment. In the proposed algorithm, both the top-view vision and depth images are captured by a video-plus-depth camera mounted on the ceiling. The depth image is first processed by a morphological operator to alleviate depth artifacts such as the optical noise and lost data. Then the human object is extracted using a human model from the preprocessed depth image. Finally, the trajectory of the detected object is established by applying the bidirectional matching algorithm. Experimental results show that the proposed algorithm achieves over 98% accuracy in various testing environments.

A hash-based RFID security protocol for strong privacy protection

Abstract: RFID (Radio Frequency Identification) tags are small, wireless electronic devices that help identify objects and people. Privacy protection and integrity assurance become rather crucial in the RFID systems, because these RFID tags may have a wide transmission range, making them subject to unauthorized scanning by malicious readers and various other attacks. Hence, Ha et al. proposed an RFID protocol and proved that their protocol can provide the forward privacy service. However, in this paper, it is shown that an attacker can track a target tag by observing unsuccessful previous session of the tag. That is, Ha et al.'s RFID protocol fails to provide the forward privacy protection as claimed. Therefore, to overcome the privacy weaknesses of Ha et al.'s RFID protocol, an RFID protocol based on the cryptographic hash functions is proposed. Moreover, the proposed RFID protocol is evaluated according to both the privacy attribute and the implementation performance.

Cooperative wireless sensor environments supporting body area networks

Abstract: Wireless Body Area Network (WBAN) in recent years have received significant attention, due to their potential for increasing efficiency in healthcare monitoring. Typical sensors used for WBAN are low powered single transceiver devices utilizing a single channel for transmission at the Medium Access Control (MAC) layer. However, performance of these devices usually degrades when the density of sensors increases. One approach to counter this performance degradation is to exploit multiple channels at the MAC layer, where optimal usage of the channels is achieved through cooperation between the sensor nodes. In this paper we propose a cooperative WBAN environment that supports multi-hop transmission through cooperation involving both environmental sensors and WBAN nodes. Our solution extends the cooperation at the MAC layer to a cross-layered gradient based routing solution that allows interaction between WBAN and environmental sensors in order to ensure data delivery from WBANs to a distant gateway. Extensive simulations for healthcare scenarios have been performed to validate the cooperation at the MAC layer, as well as the cross-layered gradient based routing. Comparisons to other cooperative multi-channel MAC and routing solutions have shown the overall performance improvement of the proposed approach evaluated in terms of packet loss, power consumption and delay.

Vehicle navigator using a mixture particle filter for inertial sensors/odometer/map data/GPS integration

Abstract: The market for vehicular navigators boomed over the last few years. These navigators rely mainly on satellite based navigation systems such as the Global Positioning System (GPS) to assist drivers. Due to interruption or degradation in such systems in dense urban scenarios, they have to be augmented with other systems to achieve continuous and accurate vehicular navigation. GPS is integrated with low-cost micro-electro mechanical system (MEMS)-based inertial sensors. However, these sensors provide inadequate performance in degraded GPS environments because of their complex error characteristics that often lead to large position drift errors. This paper proposes a continuous and accurate solution integrating low-cost MEMS-based inertial sensors, the vehicle odometer, GPS, and map data from road networks. Despite the traditional inadequate performance of MEMS-based sensors in this problem, the performance is enhanced through: (i) a special combination of inertial sensors and odometer that has better performance for land vehicles than traditional solutions; (ii) The use of map information from road networks to constrain the positioning solution; (iii) The use of an advanced particle filtering (PF) technique to perform the integration, which work with nonlinear models and better modeling of inertial sensor errors, in addition to better integration with the map data. The performance of the proposed positioning system has been verified extensively on real road tests in downtown trajectories with degraded or totally denied GPS for long durations.

Nonlinear companding transform technique for reducing PAPR of ODFM signals

Abstract: In this paper a novel nonlinear companding transform (NCT) scheme based on the inverse hyperbolic sine function is proposed for reducing the peak-to-average power ratio (PAPR) of orthogonal frequency division multiplexing (OFDM) signals. By introducing the variable companding parameters and an inflexion point in the companding function, it enables more design flexibility and freedom in companding form so that an effective trade-off between the PAPR and bit error rate (BER) performance can be offered. Moreover, the theoretical study reveals that, to approach a desired PAPR threshold, it is preferable for the proposed scheme to make the companding distortion as small as possible with appropriate selection of the companding parameters and forms. The analysis expressions regarding the achievable PAPR reduction, transform gain G, and signal attenuation factor are derived. Simulation results also demonstrate that this scheme can substantially outperform existing NCT techniques.

A survey of fast mode decision algorithms for inter-prediction and their applications to high efficiency video coding

Abstract: The emerging High Efficiency Video Coding (HEVC) standard attempts to improve the coding efficiency by a factor of two over H.264/AVC using new compression tools with high computational complexity. The increased computational complexity makes the real-time execution with reasonable computing power become one of the critical concerns for the commercialization of HEVC. A large number of prediction modes are the main causes of the increased complexity of HEVC. Thus, a fast decision of a prediction mode needs to be effectively used to reduce the computational complexity. To take advantage of large amounts of previous works and to find a guide for application to HEVC, this paper presents a survey of these efforts for the previous standards, especially for H.264/AVC, and examines the possibility of the previous algorithms to be applicable for HEVC. To this end, previous algorithms are categorized and then the effectiveness of each category for HEVC is evaluated. For this evaluation, a previous algorithm is modified for HEVC when it is not applicable to HEVC directly. Simulation results show that most previous algorithms with slight modification, in general, improve the encoding speed with a relatively small degradation of the compression efficiency. Among them, hierarchical mode decision is especially effective whereas mode pre-decision using motion or spatial homogeneity often results in inaccurate results.

Environmental audio scene and activity recognition through mobile-based crowdsourcing

Abstract: Environmental audio recognition through mobile devices is difficult because of background noise, unseen audio events, and changes in audio channel characteristics due to the phone's context, e.g., whether the phone is in the user's pocket or in his hand. We propose a crowdsourcing framework that models the combination of scene, event, and phone context to overcome these issues. The framework gathers audio data from many people and shares user-generated models through a cloud server to accurately classify unseen audio data. A Gaussian histogram is used to represent an audio clip with a small number of parameters, and a k-nearest classifier allows the easy incorporation of new training data into the system. Using the Kullback-Leibler divergence between two Gaussian histograms as the distance measure, we find that audio scenes, events, and phone context are classified with 85.2%, 77.6%, and 88.9% accuracy, respectively.

Robust video stabilization based on particle filtering with weighted feature points

Abstract: Camera global motion estimation is critical to the success of video stabilization. This paper presents an effective and robust feature based motion estimation method. In the proposed approach, feature points are collected from input video sequences based on Speeded Up Robust Features (SURF). Random Samples Consensus (RANSAC) is used to remove local motion vectors and incorrect correspondences. In the global motion estimation, a particle filter is used to estimate the weight of feature points, solving the issue of Different Depth of Field (DDOF) for feature points. Then, the weighted least square (WLS) algorithm is applied to obtain the global motion estimation. Finally, a Kalman filter estimates the intentional motion, and the unintentional motion is compensated to obtain stable video sequences. Experimental results show that the proposed algorithm has the characteristics of high precision and good robustness.