In this paper, We have developed a rolling and non-rolling subtitle detection algorithm with temporal and spatial analysis for news video. This paper mainly includes rolling subtitle region detection, character segmen...
详细信息
In this paper, We have developed a rolling and non-rolling subtitle detection algorithm with temporal and spatial analysis for news video. This paper mainly includes rolling subtitle region detection, character segmentation for rolling subtitle, and non-rolling subtitle detection and location. While, most previous video text detection methods aimed at non-rolling subtitles, this paper proposed a detailed detection algorithm for both rolling and non-rolling subtitles in currently popular news video. The proposed algorithm makes good use of typical features of rolling subtitles, and is efficient for rolling character segmentation. For non-rolling subtitles, it utilizes edge features and connected component analysis of subtitle regions. In the experiments, we have tested the proposed algorithm with a series of news video and achieved good detection and segmentation results.
During the last two decades, performance assessment of controlsystems has been receiving wide attention. However, estimation of the benchmark performance of nonlinear controlsystems still remains open. In this work,...
详细信息
During the last two decades, performance assessment of controlsystems has been receiving wide attention. However, estimation of the benchmark performance of nonlinear controlsystems still remains open. In this work, we consider an estimation problem of the benchmark performance when control loops are intervened by the complex non-differentiable nonlinearity. The considered nonlinearity includes control valve stiction of the controlsystems, as well as switching action involved by protection valves widely used in the safety-related controlsystems. Based on the idea of Lebesgue sampling, the paper proposes a novel threshold autoregressive model for unbiased estimation of the benchmark performance. Basically, the proposed method is based on the partitioning of the closed-loop routine operating data when it reaches certain thresholds. Modeling each partitioned regime as an autoregressive model, the prediction error variance as the benchmark performance can be obtained by the principle of pooled variance. A numerical example well shows the effectiveness of the proposed method.
This paper presents a method of point feature tracking and online identification using SIFT(Scale Invariant Feature Transform). The proposed approach uses the probabilistic voting method with appearance model to estim...
详细信息
This paper presents a method of point feature tracking and online identification using SIFT(Scale Invariant Feature Transform). The proposed approach uses the probabilistic voting method with appearance model to estimate the object's optimal center and apply hierarchical vocabulary tree to recognize the object. Since SIFT feature is invariant to changes caused by rotation, scaling and illumination, we can obtain higher tracking performance than the conventional approach and the probabilistic voting approach enables the track to search object efficiently. Online identification is also a challenge in video surveillance system, we use bag of words method based on hierarchical vocabulary tree to represent and match tracked objects by sampling SIFT feature online. Experimental results illustrate that the proposed approach works robustly for multi-persons tracking and identification.
In this paper we propose an image-based eyeglasses try-on method, which can be used in e-commerce. The proposed method combines an improved active shape model feature extration algorithm with an image composition algo...
详细信息
ISBN:
(纸本)9781612847719
In this paper we propose an image-based eyeglasses try-on method, which can be used in e-commerce. The proposed method combines an improved active shape model feature extration algorithm with an image composition algorithm, producing a natural facial image with eyeglasses specified by users. To preserve the color fidelity of the eyeglasses on the resulting image, a new variational model was proposed. The output image was obtained by solving the modified Poisson equations. Experimental results show that the eyeglasses image was blended into the facial image seamlessly, producing natural eyeglasses try-on effects.
For a category of linear parameter varying (LPV) systems, i.e. LPV systems with both bounded rates of parameter variations and parameter measurement errors, the approach to design the feedback robust model predictive ...
详细信息
For a category of linear parameter varying (LPV) systems, i.e. LPV systems with both bounded rates of parameter variations and parameter measurement errors, the approach to design the feedback robust model predictive control (RMPC) is studied. The proposed controller utilizes the information on system parameters so as to improve the control performance, where the LPV system model is transferred into a sequence of future models with parameter-incremental uncertainty to include both the parameter variations and the parameter measurement. Then, a sequence of feedback control laws is designed to correspond to the sequence of future models. Since the information on system parameters is utilized and the control actions will vary corresponding to the future variations of system parameters, the better control performance can be achieved. The recursive feasibility and closed-loop stability of the proposed RMPC are also proven.
Thermal efficiency of coal-fired utility boiler is an important indicator used to measure economic operation of power plant. However, on-line calculation for thermal efficiency of boiler still have difficulties, which...
详细信息
Thermal efficiency of coal-fired utility boiler is an important indicator used to measure economic operation of power plant. However, on-line calculation for thermal efficiency of boiler still have difficulties, which the coal quality are often changing and not well obtained real-time. This paper identify the heating value of the coal into the boiler, aiming at the on-line calculation of the boiler efficiency, by means of the dynamic mass and energy balance method. Then an on-line calculation model for thermal efficiency of the boiler was established based on the indirect heat balance method. The paper illustrates the effectiveness of the models as well as the validation of the on-line simulation, starting from real operation data. The outcome is that it is possible to significantly satisfy the accuracy of on-line identification of the Low Heating Value (LHV) of the coal into the boiler, and the real-time of on-line calculations provided by indirect heat balance method. The models can be used to the real-time control and optimization of the coal-fired utility boilers.
Abstract In this paper, we develop a novel self-propelled particle model to describe the emergent behavior of a group of mobile agents. All agents coordinate with their neighbors through local social forces accounting...
详细信息
Abstract In this paper, we develop a novel self-propelled particle model to describe the emergent behavior of a group of mobile agents. All agents coordinate with their neighbors through local social forces accounting for velocity alignment and collision avoidance. We allow the interaction ranges to adapt to the group density. This range adaption results in topology changes as well as discontinuities in those social forces. We apply differential inclusion technique to analyze the convergence properties of the proposed control method. The analytical and numerical results show that all the agents eventually align their velocities and avoid collisions with each other no matter how fast the topology changes.
Abstract The consensus problem for multi-agent systems with double integrator dynamics is considered in this paper. We propose and analyze a connectivity-preserving consensus algorithm with position measurements only....
详细信息
Abstract The consensus problem for multi-agent systems with double integrator dynamics is considered in this paper. We propose and analyze a connectivity-preserving consensus algorithm with position measurements only. Under the assumption that the initial interaction network is connected, the connectivity of the interaction network during the dynamical evolution can be preserved. Moreover, the consensus algorithm with a virtual leader is investigated. It is shown that all the agents can asymptotically attain a desired position and velocity even if only one agent in the team has access to the information of the virtual leader. Numerical simulations illustrate the effectiveness of the proposed methods.
Nowadays, probe vehicles equipped with Global Position system (GPS) are an effective way of collecting real-time traffic information. This paper first briefly introduces the Curve-Fitting Estimation Model (CFEM), whic...
详细信息
Nowadays, probe vehicles equipped with Global Position system (GPS) are an effective way of collecting real-time traffic information. This paper first briefly introduces the Curve-Fitting Estimation Model (CFEM), which is one of the typical methods using GPS data to estimate the traffic flow state. After that, it is detailedly analyzed how many probe vehicles the CFEM requires in order to ensure enough estimated accuracy. Furthermore, a sample size algorithm is developed to calculate the minimum sample size of the CFEM. In the algorithm, the road type, the length of road section, and sample frequency are taken into account. Finally, the proposed algorithm of sample size analysis are tested by the experiments using the data collected from the road network of the whole center region of Shanghai.
In terms of the convenience and interaction in the smart home, context awareness is introduced when we enter or leave a house resulting from that conventional methods such as RFID and pyroelectric infrared (PIR) senso...
详细信息
In terms of the convenience and interaction in the smart home, context awareness is introduced when we enter or leave a house resulting from that conventional methods such as RFID and pyroelectric infrared (PIR) sensors are usually adopted in indoor environments which are tough to determine the human motion direction. Specifically, a centroid based human motion recognition is implemented along with the eigenface based face recognition, which are both used to detect an event with the time and place added. Thus, an event driven application is constructed concerning the distinctive reactions towards different events detected. This paper initially proposes the algorithms adopted in the recognition of motion and face. Finally, with the method integration, an experiment is carried out in our demonstration room comprising the information integration of the camera videos as well as the communication in the wireless nodes that enables the reactions of the corresponding actuators, which is a positive attempt in context-aware smart home and can be further investigated afterwards.
暂无评论