Multimedia content analysis refers to the computerized understanding of semantic meanings a multimedia document, such as video sequence with an accompanying audio track. With its semantics are embedded in multiple forms that usually complimentary each other, Therefore, it is necessary analyze all types data: image frames, sound tracks, texts can be extracted from and spoken words deciphered This involves segmenting document into semantically meaningful units, classifying unit predefined...

Hyperspectral images (HSIs) are often corrupted by a mixture of several types noise during the acquisition process, e.g., Gaussian noise, impulse dead lines, stripes, etc. Such complex could degrade quality acquired HSIs, limiting precision subsequent processing. In this paper, we present novel tensor-based HSI restoration approach fully identifying intrinsic structures clean part and mixed part. Specifically, for part, use tensor Tucker decomposition to describe global correlation among all...

Rain streaks removal is an important issue of the outdoor vision system and has been recently investigated extensively. In this paper, we propose a novel tensor based video rain approach by fully considering discriminatively intrinsic characteristics clean videos, which needs neither detection nor time-consuming dictionary learning stage. specific, on one hand, are sparse smooth along raindrops direction, other videos possess smoothness rain-perpendicular direction global local correlation...

The objective of multiple description coding (MDC) is to encode a source into bitstreams supporting quality levels decoding. In this paper, we only consider the two-description case, where requirement that high-quality reconstruction should be decodable from two together, while lower, but still acceptable, reconstructions either individual bitstreams. This paper describes techniques for meeting MDC objectives in framework standard transform-based image through design pairwise correlating...

In recent studies on sparse modeling, the nonconvex regularization approaches (particularly, $L_{q}$ with $q\in(0,1)$) have been demonstrated to possess capability of gaining much benefit in sparsity-inducing and efficiency. As compared convex (say, $L_{1}$ regularization), however, convergence issue corresponding algorithms are more difficult tackle. this paper, we deal for a specific but typical scheme, $L_{1/2}$ regularization, which has successfully used many applications. More...

Information security can be compromised by leakage via low-level hardware features. One recently prominent example is cache probing attacks, which rely on timing channels created caches. We introduce a design language, SecVerilog, makes it possible to statically analyze information flow at the level. With systems built with verifiable control of and other channels. SecVerilog Verilog, extended expressive type annotations that enable precise reasoning about flow. It also comes rigorous formal...

The total variation (TV) is a powerful regularization term encoding the local smoothness prior structure underlying images. By combining TV with low rank prior, 3D (3DTV) regularizer has achieved advanced performance in general hyperspectral image (HSI) processing tasks. Intrinsically, 3DTV assumes i.i.d. sparsity structures on all bands of gradient maps calculated along spectrum and space an HSI. This, however, largely deviates from real-world cases, where generally have different while...

Background subtraction has been a fundamental and widely studied task in video analysis, with wide range of applications surveillance, teleconferencing, 3D modeling. Recently, motivated by compressive imaging, background from measurements (BSCM) is becoming an active research surveillance. In this paper, we propose novel tensor-based robust principal component analysis (TenRPCA) approach for BSCM decomposing frames into backgrounds spatial-temporal correlations foregrounds spatio-temporal...

It is known that the decomposition in low-rank and sparse matrices ( <bold xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">L+S</b> for short) can be achieved by several Robust PCA techniques. Besides low rankness, local smoothness xmlns:xlink="http://www.w3.org/1999/xlink">LSS</b> ) a vitally essential prior many real-world matrix data such as hyperspectral images surveillance videos, which makes have low-rankness property at same time. This poses an...

In this letter, we consider the problem of compressive sensing hyperspectral images (HSIs). We propose a novel tensor-based approach by modeling global spatial-spectral correlation and local smoothness properties hidden in HSIs. Specifically, use tensor Tucker decomposition to describe among all HSI bands, weighted 3-D total variation characterize smooth structure both spatial spectral modes. then design an efficient algorithm solve resulting optimization using alternating direction method...

The structural integrity of the cervix in pregnancy is necessary for carrying a until term, and organization human cervical tissue collagen likely plays an important role tissue's function. Collagen fibers extracellular matrix exhibit preferential directionality, this network ultrastructure hypothesized to reorient remodel during softening dilation at time parturition. Within cervix, upper half substantially loaded where premature funneling starts happen. To characterize we imaged whole...

Hyperspectral image (HSI) possesses three intrinsic characteristics: the global correlation across spectral domain, nonlocal self-similarity spatial and local smooth structure both domains. This paper proposes a novel tensor based approach to handle problem of HSI super-resolution by modeling such underlying characteristics. Specifically, noncovex penalty is used exploit former two characteristics hidden in several 4D tensors formed similar patches within 3D HSI. In addition, smoothness...

Hyperspectral (HS) imaging has been widely used in various real application problems. However, due to the hardware limitations, obtained HS images usually have low spatial resolution, which could obviously degrade their performance. Through fusing a resolution image with high auxiliary (e.g., multispectral, RGB or panchromatic image), so-called fusion underpinned much of recent progress enhancing image. Nonetheless, corresponding well registered cannot always be available some situations. To...

The authors demonstrate that Hopfield-type networks can find reasonable solutions to the traveling salesman problem (TSP) and optimal list-matching (LMP). They show how avoid difficulties encountered by G.V. Wilson G.S. Pawley (1988) using a modified energy functional which yields better TSP than J.J. Hopfield D.W. Tank's (1985) original formulation. In addition, two fixed-parameter are described, one for other LMP. performance of network is comparable formulation, while shown perform simple...

The low-rank tensor factorization (LRTF) technique has received increasing attention in many computer vision applications. Compared with the traditional matrix technique, it can better preserve intrinsic structure information and thus a low-dimensional subspace recovery performance. Basically, desired is recovered by minimizing least square loss between input data its factorized representation. Since most optimal when noise follows Gaussian distribution, -norm-based methods are designed to...

During pregnancy, the uterine cervix is mechanical barrier that prevents delivery of a fetus. The underlying cervical collagen ultrastructure, which influences overall properties cervix, plays role in maintaining successful pregnancy until term. Yet, not much known about this ultrastructure pregnant and nonpregnant human tissue. We used optical coherence tomography to investigate directionality dispersion fiber bundles cervix. An image analysis tool has been developed, combining stitching...