Search

• This paper proposes a new low-rank kernel regression method with preserved locality for multi-class analysis. • In this method, the regression analysis is conducted based on the low-rank components of a non-linear random feature space with preserved locality. • The impacts of the outliers hidden in the low-rank components are relieved by the l2,p norm. • The proposed method may be the first regression method that simultaneously has the abilities of non-linear analysis, redundancy reduction, and locality preservation. Kernel ridge regression (KRR) is a kind of efficient supervised algorithm for multi-class analysis. However, limited by the implicit kernel space, current KRR methods have weak abilities to deal with redundant features and hidden local structures. Thus, they may get indifferent results when applied to analyze the data with complicated components. To overcome this weakness and obtain better multi-class regression performance, we propose a new method named low-rank kernel regression with preserved locality (RLRKRR). In this method, data are mapped into an explicit feature space by using the random Fourier feature technique to discover the non-linear relationship between data samples. In addition, during the training of the regression coefficient matrix, the low-rank components of this explicit feature space are simultaneously extracted for reducing the effect of the redundancy. Moreover, the graph regularization is performed on the extracted low-rank components to preserve local structures. Furthermore, the l 2 , p norm is imposed on the regression error term for relieving the impact of outliers. Based on these strategies, RLRKRR is capable to achieve rewarding results in complicated multi-class data analysis. In the comprehensive experiments conducted on various types of datasets, RLRKRR outperforms several state-of-the-art regression methods in terms of classification accuracy (CA). [ABSTRACT FROM AUTHOR]

Abstract: Background: Screws with strong pull-out strength have been sought for the treatment of cancellous bone. We hypothesized that an obliquely angled screw thread has advantages over conventional vertical thread with a minimal proximal half angle. Methods: Metal and bone screws were made of stainless steel and porcine cortical bone. Their proximal half angle was set at 0°, 30°, or 60°. The screws were inserted into porcine cancellous bone. At 0°, the thread faced the recipient bone vertically. Pullout tests at a rate of 30mm/min (n =40, each screw type) and microcomputed tomography (n =6) were conducted. Findings: The pull-out strength of the screws was maximal at 30°; 348.8 (SD, 44.1)N with metal and 326.6 (39.4)N with bone. It was intermediate at 0°; 301.9 (35.9)N with metal and 278.2 (30.6)N with bone. It was minimal at 60°; 126.5 (39.0)N with metal and 174.8 (29.7)N with bone. Cancellous bone was damaged between the threads at 30°, while intact cancellous bone was preserved between the threads at 0°. Interpretation: A proximal half angle of around 30° is appropriate because the pullout force is applied to the recipient bone evenly. Commercial cancellous screws can be improved by changing the thread shape to minimize the damage to recipient bone. [Copyright &y& Elsevier]

Schwertmannite (Sch) is known to be an effective scavenger of arsenic (As) due to its strong binding affinity for toxic As species. However, the evolution of As-bearing schwertmannite under extreme pH conditions is poorly understood. In this study, we investigated the effects of extremely acidic and alkaline conditions on the stability of schwertmannite with structurally incorporated As(V) (CO-Sch) and schwertmannite with adsorbed As(V) (AD-Sch). The results show that both extremely acidic and alkaline conditions have significant effects on the evolution of minerals and liberation of iron and sulfate. At extremely acidic pH, the maximal release of ferric iron (Fe(III)) and sulfate from CO-Sch were greater than that from AD-Sch, whereas 6.2% and 0.3% of total As released from AD-Sch and CO-Sch, respectively. At extremely alkaline pH, aqueous Fe(III) was not observed, and Fe(III) was retained in As-bearing schwertmannite due to the chemical equilibrium between the dissolution of schwertmannite and re-precipitation of goethite; structurally incorporated As(V) promoted the liberation of sulfate. In addition, the adsorbed As on schwertmannite is more stable, which led to a minor release of As (0.8%) over a 30-d period, however, the liberated As(V) from CO-Sch accounts for up to 3.2%. Under extremely acidic and alkaline conditions, portions of AD-Sch and CO-Sch transformed from schwertmannite to goethite after 30 d, while schwertmannite was still the dominant mineral. Adsorbed As(V) inhibited the transformation of As-bearing schwertmannite to goethite more significantly than structurally incorporated As(V). • Extreme acidic pH favored the transformation of As-bearing schwertmannites. • pH and As(V) speciation influenced the release of iron and sulfate significantly. • As(V) speciation affected differently the mobility of As at extreme pH conditions. [ABSTRACT FROM AUTHOR]

Image clustering is a difficult task with important application value in computer vision. The key to this task is the quality of images features. Most of current clustering methods encounter the challenge. That is, the process of feature learning and clustering operates independently. To address this problem, several researchers have been dedicated to performing feature learning and deep clustering together. However, the obtained features lack discriminability to address high-dimensional data successfully. To deal with this issue, we propose a novel model named as robust deep fuzzy K -means clustering (RD-FKC), which efficiently projects image samples into a representative embedding space and precisely learns membership degrees into a combined framework. Specifically, RD-FKC introduces Laplacian regularization technique to preserve locality properties of data. Moreover, by using an adaptive loss function, the model becomes more robust to diverse types of outliers. Furthermore, to avoid the latent space being distorted and make the extracted features retain the original information as much as possible, the model introduces reconstruction error and adds regularization to network parameters. Finally, an effective algorithm is derived to solve the optimization model. Numerous experiments have been conducted, illustrating the advantages and superiority of RD-FKC over existing clustering approaches. • Embedding clustering into deep convolutional autoencoder to learn compact and representative features. • Designing the Laplacian regularization to constrain the membership matrix for locality preserving. • Introducing the adaptive loss function to enhance the clustering robustness. • Presenting an effective algorithm to optimize the proposed model. [ABSTRACT FROM AUTHOR]

Bone screws have a mechanical advantage over bone pegs for fixation of osteochondral fractures. The purpose of this study was to investigate predrilling methods to facilitate bone screw insertion. The 162 bone screws (major diameter 3.0 mm; minor diameter 2.5 mm; length 10 mm) used were made from porcine tibial cortical bone. Metal screws of the same design were used as a control. The screws were inserted, without tapping, into predrilled holes of six different sizes (2.4–2.9 mm diameter; 80%–97% of the screw) made in cancellous bone harvested from porcine femoral condyle, and pullout strength was determined. The volumetric bone mineral densities of porcine bones were compared to those of human bones. In a clinical case, an osteochondral fracture of the patella was repaired using three autologous bone screws. Bone screws occasionally broke due to high insertion torque (108–158 N·mm); the incidences were 30% and 19% with the 80% and 83% sized holes, respectively. None of the screws inserted into holes of size ≥ 87% broke, and their insertion torques never exceeded 100 N·mm. Metal screws were inserted without breakage regardless of hole size. Bone screws inserted into 80%-93% holes had the greatest pullout strengths (range 154–165 N), whereas 97% holes had a mean strength of 84 N. Metal screws had large pullout strengths when inserted into 80% and 83% holes (164 and 169 N, respectively). The mineral density of porcine bone was lower than that of human bone (1167 and 1357 mg/cm3 for cortical bone and 193 and 269 mg/cm3 for cancellous bone, respectively). In the clinical case, bone screws were inserted smoothly into 90% holes. The optimal size of predrilled holes for bone screws (87%–93%) differs from that for metal screws. [ABSTRACT FROM AUTHOR]

In image set classification, affine hull based models have achieved good performance. However, these models only focus on holistic images and ignore the local information of images. To incorporate local information, this paper develops a cooperative operator and proposes a cooperative linear regression model (CLRM) for image set classification. Specifically, CLRM first proposes a block partitioning strategy to obtain local information. Image sets and local information are both modeled as affine hulls. Then, a linear regression model embedded with cooperative operator is proposed to incorporate different information and compute the distance between the different affine hulls. Finally, a combined distance measure is proposed to incorporate different information for classification. Extensive experiments on different recognition tasks such as face recognition on image/video based databases and object classification show that the proposed cooperative operator can improve the performance of the model and the proposed CLRM achieves competitive performance compared with the state-of-the-art classifiers. • Proposing a block partitioning strategy to extract local information. • Modeling local information as affine hull for image set classification. • Developing a cooperative operator to integrate different information as a whole. • Proposing a linear regression model embedded with the cooperative operator. • Proposing an efficient optimization algorithm to solve the model. [ABSTRACT FROM AUTHOR]

Intestinal mucosal barrier damage is an important mechanism for the development of sepsis and multiple organ dysfunction syndrome. At present, there are no satisfactory and effective methods for the protection of the intestinal mucosal barrier. Jinzhi, the first fecal microbiota transplantation worldwide, is often used to treat critically ill patients; however, the specific mechanism involved in this process remains unclear. The aim of this study was to investigate the therapeutic effect and mechanism of Jinzhi intervention on mice with sepsis induced through treatment with lipopolysaccharide (LPS). Mice were intraperitoneally injected with LPS to simulate intestinal mucosal barrier function damage in sepsis; intervention was performed through the oral administration of Jinzhi. The effect of Jinzhi on LPS-induced sepsis was analyzed by comparing the vital signs and survival rate of mice under different treatments. Pathological staining and enzyme-linked immunosorbent assay were used to identify the effects of LPS or treatment with Jinzhi on the intestinal mucosal barrier in mice. The effect of LPS or treatment with Jinzhi on the intestinal flora was analyzed via 16S rRNA gene sequencing of ileal contents. Immunohistochemistry and enzyme-linked immunosorbent assay showed that treatment with LPS increased levels of inflammatory factors (interleukin-1α, interleukin-6, tumor necrosis factor-α), caspase-3, and caspase-8 in the serum and ileum, and destroyed the tight junction between epithelial cells. Intervention with Jinzhi reduced levels of serum LPS and tumor necrosis factor-α, and repaired the tight junction between epithelial cells. Furthermore, 16S rRNA gene sequencing analysis showed that treatment with Jinzhi improved the diversity and physiological function of the intestinal flora. These results suggest that Jinzhi may be a promising option for the treatment of sepsis caused by LPS, and emphasize that Jinzhi exerts a recovery effect on the imbalance of intestinal flora. [ABSTRACT FROM AUTHOR]