内容一致性行人重识别算法

行人重识别是指利用计算机视觉技术识别不同监控设备下的目标行人,该技术在公共安全与相册管理等方面应用较广。然而现有行人重识别算法在局部特征区域划分后出现离异值使该区域内容不一致,导致局部特征可区分性降低。提出一种基于局部区域特征选择的内容一致性行人重识别算法。将行人图像输入残差卷积神经网络取得张量,根据局部区域内容一致性从张量中选择基本单位特征向量,使用Softmax函数计算其局部区域概率重新生成局部区域,从而消除离异值,增加类间差异并减少类内差异。实验结果表明,与Spindel、PN-GAN等行人重识别算法相比,该算法的行人重识别准确率更高,其提取的行人特征可区分性和鲁棒性更好。

A blind watermarking algorithm based on DWT and SVD

This paper presents a new digital image blind watermarking algorithm based on combination of discrete wavelet transform （DWT） and singular value decomposition （SVD）. First of all, we make wavelet decomposition for the original image and divide the acquired low frequency sub-band into blocks. Then we make singular value decomposition for each block and embed the watermark information in the largest singular value by quantitative method. The watermark can be extracted without the original image. The experimental results show that the algorithm has a good imperceptibility and robustness.

Audio Zero-Watermark Scheme Based on Discrete Cosine Transform-Discrete Wavelet TransformSingular Value Decomposition

Traditional watermark embedding schemes inevitably modify the data in a host audio signal and lead to the degradation of the host signal.In this paper,a novel audio zero-watermarking algorithm based on discrete wavelet transform(DWT),discrete cosine transform(DCT),and singular value decomposition(SVD) is presented.The watermark is registered by performing SVD on the coefficients generated through DWT and DCT to avoid data modification and host signal degradation.Simulation results show that the proposed zero-watermarking algorithm is strongly robust to common signal processing methods such as requantization,MP3 compression,resampling,addition of white Gaussian noise,and low-pass filtering.

A new algorithm for estimating gillnet selectivity

The estimation of gear selectivity is a critical issue in fishery stock assessment and management.Several methods have been developed for estimating gillnet selectivity,but they all have their limitations,such as inappropriate objective function in data fitting,lack of unique estimates due to the difficulty in finding global minima in minimization,biased estimates due to outliers,and estimations of selectivity being influenced by the predetermined selectivity functions.In this study,we develop a new algorithm that can overcome the above-mentioned problems in estimating the gillnet selectivity.The proposed algorithms include minimizing the sum of squared vertical distances between two adjacent points and minimizing the weighted sum of squared vertical distances between two adjacent points in the presence of outliers.According to the estimated gillnet selectivity curve,the selectivity function can also be determined.This study suggests that the proposed algorithm is not sensitive to outliers in selectivity data and improves on the previous methods in estimating gillnet selectivity and relative population density of fish when a gillnet is used as a sampling tool.We suggest the proposed approach be used in estimating gillnet selectivity.

Some results on the regularization of LSQR for large-scale discrete ill-posed problems

LSQR, a Lanczos bidiagonalization based Krylov subspace iterative method, and its mathematically equivalent conjugate gradient for least squares problems(CGLS) applied to normal equations system, are commonly used for large-scale discrete ill-posed problems. It is well known that LSQR and CGLS have regularizing effects, where the number of iterations plays the role of the regularization parameter. However, it has long been unknown whether the regularizing effects are good enough to find best possible regularized solutions. Here a best possible regularized solution means that it is at least as accurate as the best regularized solution obtained by the truncated singular value decomposition(TSVD) method. We establish bounds for the distance between the k-dimensional Krylov subspace and the k-dimensional dominant right singular space. They show that the Krylov subspace captures the dominant right singular space better for severely and moderately ill-posed problems than for mildly ill-posed problems. Our general conclusions are that LSQR has better regularizing effects for the first two kinds of problems than for the third kind, and a hybrid LSQR with additional regularization is generally needed for mildly ill-posed problems. Exploiting the established bounds, we derive an estimate for the accuracy of the rank k approximation generated by Lanczos bidiagonalization. Numerical experiments illustrate that the regularizing effects of LSQR are good enough to compute best possible regularized solutions for severely and moderately ill-posed problems, stronger than our theory, but they are not for mildly ill-posed problems and additional regularization is needed.