Person reidentification (ReID) is a very popular research topic in machine learning and computer vision. According to the occlusions, it can be divided into holistic person ReID partial tasks. Occlusions commonly exist task but pose or observation perspective changes often occur task; thus, many different algorithms network architectures have been designed for each task. However, this approach increases cost practice hinders development of techniques. To solve problem, work, unified...

We demonstrate a novel high speed and multi-bit optical quantum random number generator by continuously measuring arrival time of photons with common starting point. To obtain the unbiased post-processing free bits, measured photon is converted into sum integral multiple fixed period phase time. Theoretical experimental results show that an independent uniform variable. A bit extraction method encoding proposed. An setup has been built generation rate could reach 128 Mb/s, efficiency 8 bits...

The directional total variation (DTV) model has been proposed very recently for image denoising. However, the DTV works well when there is just one dominant direction in image. In this Letter, authors propose to make adaptive edge so that can handle images with several directions. Experiment and comparison show effectiveness of method.

In motor-imagery brain–computer interface (BCI), transfer learning based on the framework of regularized common spatial patterns (RCSP) can make full use training data derived from other subjects to reduce calibration time for a new subject. Covariance matrices are commonly used estimate difference between subjects. However, classification performances vary greatly depending different assumptions distribution covariance matrices. Therefore, directly observe variations target subject’s...

In this paper, human action recognition using pyramid histograms of oriented gradients and collaborative multi-task learning is proposed.First, we accumulate global activities construct motion history image (MHI) for both RGB depth channels respectively to encode the dynamics one in different modalities, then descriptors are extracted from MHI represent textual structural characteristics these actions.Specially, average value hierarchical block, GIST employed motion.To demonstrate...

To solve the problem of extracting visual feature in lipreading, a new method based on DCT+LDA is proposed this paper. First, region interest (ROI) located lip contour information, and then discrete cosine transformation (DCT) performed ROI. In order to extract most discriminative vectors from DCT coefficients further reduce dimensionality, linear analysis (LDA) introduced. Experiments were speaker-dependent (SD) speaker-independent (SI) bimodal database respectively, experimental results...

The traditional predictive methods for tracking moving objects usually assume that have linear motion patterns. This severely limits their applicability, since in practice movement is free and uncertain. In this paper, a novel location prediction model based on grey theory presented. proposed adopts the modeling method to predict future of uncertain objects. Comparing with model, relaxes limitation pattern requirement accuracy sampling data. experiment results show can provide more exact than model.

Long and tedious calibration time hinders the development of motor imagery- (MI-) based brain-computer interface (BCI). To tackle this problem, we use a limited labelled set relatively large unlabelled from same subject for training on transductive support vector machine (TSVM) framework. We first introduce an improved TSVM (ITSVM) method, in which comprehensive feature each sample consists its common spatial patterns (CSP) geometric feature. Moreover, concave-convex procedure (CCCP) to...

This paper aims to study the corrosion resistance and interfacial bonding properties of silver zinc-rich coating/substrate system silicone system. The actual working state in marine environment was simulated by coupled effects pre-stretching. Scanning electron microscope (SEM) with an energy dispersive X-ray spectrometer (EDS) used test analyze surface morphology products. Meanwhile, degradation degree investigated based on electrochemical impedance spectroscopy (EIS) water absorption rate....

The image reconstruction process in super-resolution structured illumination microscopy (SIM) is investigated. pattern generated by the interference of two Gaussian beams to encode undetectable spectra into detectable region microscope. After parameters estimation pattern, encoded are computationally decoded and recombined Fourier domain equivalently increase cut-off frequency microscope, resulting extension a reconstructed with about two-fold enhanced resolution. Three different methods...

With the development of 3D technologies and devices, model retrieval becomes a hot research topic where multi-view matching algorithms have demonstrated satisfying performance. However, exciting works overlook common factors among objects in single class, they are time consuming processing. In this paper, class-statistics pair-constraint (CSPC) method is originally proposed for retrieval, which composed supervised class-based statistics object model. our CSPC model, we firstly convert...

A weighted Sobel operator-based no-reference perceptual blockiness metric (WSBM) which is used to measure blocking artifacts of DCT coding image, presented in this paper. The proposed calculates the image luminance gradient matrix using operator. Based on matrix, it quantifies severity and annoyance large flatness low bit-rate images. new approach also takes into account spatial human visual system (HVS) features, include adaptation texture masking. These processes are individually conducted...

MeshStereo (MS) and cross‐scale cost filtering (CSCF) are two most recently celebrated models for stereo matching. On one hand, MS model enlightens fast solving the dense correspondence problem according to a region‐based opinion. other CSCF could generate more robust matching volumes than single scale. In this study, authors weave these together attaining greater faster disparity estimation. With CSCF, powerful initial of computed they conducted as data term energy function model. More...

The Chan-Vese model is one of the most popular region-based active contours, and its vector-valued extension also powerful for multichannel images. Very recently, histogram introduced into due to effectiveness region information. Motivated by fact that a tool characterize texture, it texture segmentation in this work. In order determine an optimal number bins histogram, Bayesian method adopted. Experiments are conducted results show proposed strategy effective segmentation.

The human visual system (HVS), affected by viewing distance when perceiving the stereo image information, is of great significance to study stereoscopic quality assessment. Many methods assessment do not have comprehensive consideration for perception characteristics. In accordance with this, we propose a Rich Structural Index (RSI) Stereoscopic Image objective Quality Assessment (SIQA) method based on multi-scale To begin with, put pair into pyramid Contrast Sensitivity Function (CSF)...

Since Otsu algorithm does not take the image spatial neighbor information into consideration, we combine Markov random field with to integrate gray level and correlation for pixels. In this paper, thresholding based on Random Field is proposed. algorithm, neighborhood rejectability function imported an threshold selection improved. The experiment results verify that applying our road segmentation can achieve good effects.