Search

Proposed a block-based compressive video sensing algorithm, which allows to pursue video acquisition and video compressed coding synchronously. To explore the temporal redundancy of the video, different sensing strategies were used between reference frames and non-reference frames: for reference frames, the frames were divided to little patches and employed regular compressive sensing to every patch; while for the rest ones, first the frams were divided into blocks in the same size and then compared the blocks with the corresponding block in reference frame, pursue different sensing method according to the results. The frame quality is better because non-reference frame observation could feedback the reference frames. Meanwhile, the sampling rate changes adaptively according to the texture complication. The experimental results with 20% less samples than other methods indicate that the algorithm is more suitable to human eyes and also gets higher PSNR. [ABSTRACT FROM AUTHOR]

The local linear embedding relationship of pixels was used as prior to constrain the reconstruction of HR facial image. The proposed algorithm selected K neighboring face samples, which were more similar to the input face, from the training set, and then took them as reference examples after their registrations so as to learn the embedding coefficients of pixels in target HR image. The weights of respective face examples in the learning process were adaptively adjusted in order to reduce the influence of registration errors, and the total variation minimization was used to constrain the smoothness of embedding coefficients. Combining the local pixel embedding relationship and the degradation model, the method could reconstruct target HR image using the maximum posterior estimation framework. Experimental result shows that the reconstructed HR images can posses more delicate and clear local features, and the PSNR and SSIM are 1.26 dB and 0.04 higher than that of using comparison algorithms, respectively. [ABSTRACT FROM AUTHOR]

In this paper, a human face tracking algorithm combing color and texture features based on particle filter is proposed. The proposed approach makes use of the characteristics of particle filter which not only can effectively deal with nonlinear and non Gaussian process but also can combine multiple face features information. Different robustness to different environments of features is considered and the weighted color histogram and rotated complex wavelet filter (RCWF) are used to describe features. Thus the color and texture features can be fused under particle filter framework to develop the new face tracking algorithm. Experimental results demonstrate that, compared with the method based on single feature, the robustness, accuracy and flexibility of the algorithm have the better effectiveness of the performance, and the experiments also show the accuracy for face tracking in real scenes. [ABSTRACT FROM AUTHOR]

Presents a particle filter-based algorithm for IR target-tracking. In this algorithm, sampling particles are optimized and selected, and the re-sampling phase is improved. It solves the problems of massive computation in traditional particle filter. And this method is adopted in tracking pupil, which has better precision and higher efficiency. The paper also applied Hough transform to locate the edge o{ pupil and achieved very good results. [ABSTRACT FROM AUTHOR]