Nowadays, surface defect detection systems for steel strip have replaced traditional artificial inspection systems, and automatic offer good performance when the sample set is large model stable. However, trained does work well a new production line initiated with different equipment, processes, or devices. These variables make just tiny changes to real-world but significant impact on classification result. To overcome these problems, we propose an evolutionary classifier Bayes kernel (BYEC)...

Objective:To compare the changes in distortion product otoacoustic emission （DPOAE） test results clinical patients with negative middle ear pressure after equalizing external canal and cavity. This study aims to analyze effect of on emissions investigate correlation between degree amplitude signal-to-noise ratio DPOAE. Methods:Twenty-seven were included, 34 ears exhibiting pressure. Acoustic conductance tests, pure tone hearing threshold DPOAE tests conducted under ambient peak for all...

Recently, the Coherent Point Drift (CPD) algorithm has become a very popular and efficient method for point set registration. However, this does not take into consideration neighborhood structure information of points to find correspondence requires manual assignment outlier ratio. Therefore, CPD is robust large degrees degradation. In paper, an improved proposed overcome two limitations CPD. A descriptor, such as shape context, used perform auxiliary calculation correspondence, proportion...

Alzheimer's disease (AD) has raised extensive concern in healthcare and academia as one of the most prevalent health threats to elderly. Due irreversible nature AD, early accurate diagnoses are significant for effective prevention treatment. However, diverse clinical symptoms limited neuroimaging accuracy make challenging. In this article, we built a brain network each subject, which assembles several commonly used data simply reasonably, including structural magnetic resonance imaging...

In this letter, a novel method based on discrete space model (DSM) and support vector machines (SVMs) is proposed for hyperspectral image (HSI) classification. The DSM approach transforms continuous spectral signatures into features constructs with the features. Therefore, classification capability of SVMs can be improved account feature space. Moreover, composite kernel employed to take advantage spatial among neighboring pixels. applied real HSIs experimental results confirm that accuracy...

A band selection algorithm based on information entropy is proposed for hyperspectral image classification. First, original spectral features are transformed into discrete and represented by a space model. Then, the adopted to reduce feature dimensionality. The bands with weak class separability effectively abandoned algorithm. Moreover, support vector machine classifiers composite kernels employed incorporate spatial features, reducing speckle errors in classification maps. methods applied...

Graft‐versus‐host disease (GVHD) and infection are major obstacles to successful allogeneic bone marrow transplantation (allo‐BMT). In an attempt improve the results of HLA‐identical sibling BMT, we investigated effect accelerating hemopoietic reconstitution reducing acute GVHD (aGVHD) in allo‐BMT receiving G‐CSF‐stimulated donor preliminary biological mechanism. The donors 30 patients (study group) with leukemia were given G‐CSF 3–4 μg/kg/d for 7 doses prior harvest. subsequent engraftment...

In this paper, a content adaptive coding method for plenoptic video is proposed to reduce both of the spatial and temporal redundancy. Based on correlations among subapertures, divided into two subaperture groups: central view videos residual videos. Coding methods, multiview based correlation(S-MVC) correlation(T-MVC), are adopted according correlation energy analysis, while always encoded by S-MVC. Correlation analysis performed first, further if necessary. separately reduces bitstream,...

Image completion techniques are required to fill in missing regions digital images. We propose a method that focuses on the complex structure of foreground object. observe if we match complete deformable contour sample object and obtain (i.e., point set positions), damaged can be reconstructed. further find reconstructed provides sufficient information image, which generates guidance maps as soft constraints. With maps, then incomplete image parts by combining textures both using...

Surface defects detection system has been receiving increased attention as its precision, speed and less cost. One of the most challenges is reacting to accuracy deterioration with time aged equipment changed processes. These variables will make a tiny change real world model but big impact on classification result. In this paper, we propose new adaptive classifier Bayes kernel (BYEC) which update small sample it for deterioration. Firstly, abundant features were introduced cover lots...

In this paper, we propose an improved detail enhancement method via guided image. To achieve the best overall balance for contradictory goals of edge-preserving smoothing and details capturing, a combining image filter (GIF) with local weighted least squares (WLS) global intensity shift input do so, first make operators in R, G, B channels respectively by filter. We then utilize former enhanced as model execute luminance channel. The important contents are well preserved without distorting...

Automatic image completion techniques have difficulty processing images in which the target region has multiple planes or is non-facade. Here, we propose a new method that uses belief propagation based on planar priorities. We first calculate information, includes projection parameters, plane segments, and repetitive regularity extractions within plane. Next, convert this information into guide knowledge using prior probabilities of patch transforms offsets. Using energy discrete Markov...

Image completion is aim to fill the missing regions in images.A robust technique using similarity analysis and transformation proposed address this problem.Firstly, order decrease search space of patches, random mapping method used analyze texture which have similar structure with damaged regions.Secondly, geometric photometric transformations image are adopted find best patches.Thirdly, increasing accuracy propagation, a priority calculation optimized based on confidence factor edge...

The dense connection is a powerful technique to build wider and deeper convolution neural networks (CNNs) for handling several computer vision tasks. Despite the excellent performance, it consumes numerous parameters produces large weight model file. We studied distribution of layers proposed hard parameter sharing approach known as pool (CP) compressing dense-connection-based image restoration CNN models. CP used reallocate specific ensure that some can be shared in different layers. design...

In the practical application, non-rigid point set registration should be robust for noise, occlusion or outliers. this paper, Gaussian Mixture Model (GMM) and neighborhood structure information are used registration. is to represent model set, transformation built by using radial basis function. The proportion of each component decided points. E-step EM algorithm correspondence solved, in M-step outlier ratio closed-form solution calculated. Until convergence optimal obtained. As compared...

Introduction The study is to assess outcomes of patients undergoing reconstruction with the anter-olateral thigh flap vascularized fascia lata (ALT-FL) in subsequent oncologic resection for carcinoma oral cavity, oropharynx, and maxillary, as well osteoradionecrosis debridement skull base.

Abstract Image completion is an approach to fill a damaged region (hole) in image. In this study, we adopt novel method which can repair target with structural constraints architectural scene. An objective function that consists of three terms proposed solve the image problem. color term, compute parameterized transformation model using detected plane parameters and measure distance between patch transformed source patch. This helps extend search space find optimal solution. To improve...

We aim to evaluate the efficacy and safety of basiliximab in liver transplantation (LT) for patients with hepatitis B virus-related diseases.A total 268 diseases undergoing LT were enrolled divided into two groups according usage basiliximab. Total survival, survival high-risk defined by posttransplant model predicting mortality, acute rejection rate, biochemical parameters other follow-up data compared between groups.Group Bas was composed 131 who received basiliximab, Group Triple 137 did...

Image completion techniques are required to complete missing regions in digital images. A key challenge for image is keeping consistency of structures without ambiguity and visual artifacts. We propose a novel method using depth cue. Our includes three major features. First, we compute the gradient improve when searching most similar patches. Second, depth, guide by means appropriate scale transformation. Third, global optimization patch-based having features completion. Experiments...

Image completion is an approach to repair a damaged region (a hole) in image. In this study, novel image algorithm proposed using texture direction guide the process. We first analyse and detect dominant gray-level co-occurrence matrix (GLCM). Based on detected direction, we then generate map that serves as soft constraint during Further, propose objective function can balance global structure local coherency order maintain consistency. Experimental results show method has potential...