Blind quality assessment for screen content images via convolutional neural network

With the wide propagation of cloud and mobile computing, screen content images (SCIs) have become more indispensable in our daily lives. Compared to natural scene images (NSIs), SCIs possess many particular characteristics, like mixed contents, extremely sharp edges, and text graphics. Consequently, more challenges occur in the feature extraction, which is used to reflect the distortion, during the quality assessment of SCIs. Recently, some convolutional neural network (CNN) models have been designed by automatically learning feature to evaluate the quality. In this paper, we develop a novel blind quality assessment method for SCIs via the CNN. First, compared with existing CNN-based methods, the proposed method avoids the disadvantage of training with image patches, and it is the pioneering attempt that takes the entire image as inputs. Second, instead of the image gray value, the original image is decomposed into two portions, i.e., the predicted and unpredicted portions, according to the internal generative mechanism (IGM) theory as the input of CNN. Through the CNN, all features of the image are learned automatically from beginning to end, and the network finally outputs the predicted score. Since existing SCI database is too small, to fully train the network, we collected 30000 SCIs and employed a high-accuracy full-reference quality assessment metric of SCI to compute scores as the training labels. Experimental results on SIQAD database demonstrate that the proposed method is comparable to reference-based SCI quality assessment metrics and is superior to the state-of-the-art NSI quality assessment metrics.     

Generating video animation from single still image in social media based on intelligent computing

Bringing a single still image into reality is a challenging topic in computer animation because the driven and structural information in single still image is inadequate. In this paper, we present an image animating method for enhancing single still image in social media with virtual realistic and animated motions without prior information. We imitate the interaction between the active objects in an image and their neighboring passive objects. The existing actions in the image and the virtual specified force are employed to animate the active objects. Observing that the change between two subsequent motions of the active objects derives a motion tendency, we can calculate a virtual driving force based on the motion tendency. By virtue of the virtual driving force, the stochastic motion texture is used to animate the passive objects. Finally, the convolutional neural network is employed to optimize the virtual motion animations. In this way, the proposed method produces visually natural results while guaranteeing motion harmony between active objects and passive objects. To demonstrate the applicability and rationality of virtual animation driving force, our method generates several animations from still images in Social Media.     

Facile joining of SiC ceramics with screen-printed polycarbosilane without pressure

Joining of SiC ceramics was successfully achieved at a relatively low temperature of 1500 °C without any pressure using pure polycarbosilane (PCS) as the joining material, which was distributed homogenously on the surface of SiC monolith through a screen printing method. The XRD pattern shows that the pyrolysis product of PCS is single-phase SiC. The interlayer thickness of the SiC joint is approximately 2 μm. This ultra-thin interlayer with lower possibility of the existence of defects contributes to the average shear strength of 105.8 ± 10.4 MPa, higher than that of other works using other preceramic polymers to the best of our knowledge. Due to the simplicity, low cost and high joining strength, the screen printing method using PCS as the joining material has good practicality in SiC ceramics joining.     

Stereo-camera-based object detection using fuzzy color histograms and a fuzzy classifier with depth and shape estimations

This paper proposes a new method of detecting an object containing multiple colors with non-homogeneous distributions in complex backgrounds and subsequently estimating the depth and shape of the object using a stereo camera. To extract features for object detection, this paper proposes fuzzy color histograms (FCHs) based on the self-splitting clustering (SSC) of the hue-saturation (HS) color space. For each scanning window in a pyramid of scaled images, the FCH is obtained by accumulating the fuzzy degrees of all of the pixels belonging to each cluster. The FCH is fed to a fuzzy classifier to detect an object in the left image captured by the stereo camera. To find the matched object region in the right image, the left and right images are first segmented using the SSC-partitioned HS space. The depth of the object is then found by performing stereo matching on the segmented images. To find the shape of the object, a disparity map is built using the estimated object depth to automatically determine the stereo matching window size and disparity search range. Finally, the shape of the object is segmented from the disparity map. The experimental results of the detection of different objects with depth and shape estimations are used to verify the performance of the proposed method. Comparisons with different detection and disparity map construction methods are performed to demonstrate the advantage of the proposed method.     

Statistical Analysis for Decision Making

In this article we consider a generalization of the univariate g-and-h distribution to the multivariate situation with the aim of providing a flexible family of multivariate distributions that incorporate skewness and kurtosis. The approach is to modify the underlying random variables and their quantiles, directly giving rise to a family of distributions in which the quantiles rather than the densities are the foci of attention. Using the ideas of multivariate quantiles, we show how to fit multivariate data to our multivariate g-and-h distribution. This provides a more flexible family than the skew-normal and skew-elliptical distributions when quantiles are of principal interest. Unlike those families, the distribution of quadratic forms from the multivariate g-and-h distribution depends on the underlying skewness. We illustrate our methods on Australian athletes data, as well as on some wind speed data from the northwest Pacific.     

A Comparison of Three Types of Manual Controls on a Third-Order Tracking Task

Abstract

Three types of control levers— the spring-centrod-displaccment control, the on-off control and the pressure control— were compared on a third-order compensatory tracking task. Based on an analysis of the control transfer functions, the predicted ranking of the three controls is: (1) pressure, (2) displacement and (3) on-off. The results of a study in which five subjects tracked six trials with each control for 18 sessions were as follows. The data at the beginning of training indicated rankings of (1) pressure, (2) on-off and (3) displacement. However, the data taken during the latter sessions ranked the controls as predicted, although the difference between displacement and on-off was not statistically significant. Differences in results between, early and late training, and comparisons of these results with other studies, are discussed.     

Shape optimization of cold-formed steel columns with fabrication and geometric end-use constraints

The objective of this paper is to present constrained optimization results for a cold-formed steel (CFS) cross-section shape with maximum axial capacity. In the authors׳ previous work unconstrained shape optimization was performed via stochastic search and gradient-based algorithms. Unconstrained shape optimizations produced a significant capacity increase, more than 140%, above standard CFS cross-sections, but many of the solutions are highly unconventional and have potential limitations both with respect to end use (e.g. attaching boards for walls and floors) and cost of manufacturing. Column capacity is determined using the Direct Strength Method (DSM) which requires inputs for the local, distortional and global critical buckling loads. These critical loads are obtained using the finite strip method, as implemented in the open source software CUFSM, which allows essentially any potential cross-section to be evaluated. To advance the applicability of the optimized results, end-use constraints and manufacturing constraints on the number of rollers employed in forming were both successfully incorporated in the shape optimization presented in this paper, resulting in optimized cross-sections that are more practical and economical with only marginally decreased capacity (usually less than 10%) from the earlier unconstrained optimized solutions. The constraints are implemented within a simulated annealing (SA) algorithm for the optimization. Optimized sections from multiple runs show uniformity, partially indicating the robustness of the final optimized shapes. The implemented constrained shape optimization provides a thorough search with high computational efficiency. The optimized cross-sections from this research provide promising potential shapes for the development of new commercial product families, and the member-level optimization methodology can also be integrated into building optimization in the future.     

Measurement of Bulk Mechanical Properties and Modeling the Load-Response of Rootzone Sands. Part 1: Round and Angular Monosize and Binary Mixtures

The bulk mechanical properties of two different types of rootzone sands (round and angular) were measured using a cubical triaxial tester. Two monosize sands (d 50 = 0.375 mm and 0.675 mm) and their 50:50 binary mixtures (d 50 = 0.500 mm) were studied. The compression, shear, and failure responses of the above-mentioned six compositions were analyzed, compared, and modeled. Two elastic parameters (bulk and shear moduli) and two elastoplastic parameters (swelling and consolidation indices) of the six sand compositions were also calculated and compared. The angular sand was more compressible than round sand during isotropic compression. In addition, the angular sands tended to have lower initial bulk density and high porosity values. Among the three different size fractions, the 0.375 mm mixture was least compressible for both sand shapes. The failure strength and shear modulus of the angular sand were higher than the round sands. In addition, due to their simplicity, phenomenological models were developed to predict the compression and shear behavior of the sands. The prediction models were validated using subangular and subround sands. Average relative difference values were calculated to determine the effectiveness of the prediction models. The mean average relative difference values for compression profiles, i.e., volumetric stress vs. volumetric strain, were from 16 % to 39 %, except for the initial load-response portion (< 1 % volumetric strain). The predictive models were effective in reproducing the failure responses: at 17.2 kPa confining pressure, the mean of average relative difference was 23 %; at 34.5 kPa , the mean difference was 24 %.     

Hedonic mediation of the crossmodal correspondence between taste and shape

Crossmodal correspondences between gustatory (taste), olfactory (smell), and flavour stimuli on the one hand and visual attributes on the other have been extensively documented in recent years. For instance, people have been shown to consistently match specific tastes and flavours to particular visual shapes. That said, further research is still needed in order to clarify how and why such correspondences exist. Here, we report a series of four experiments designed to assess what drives people’s matching of visual roundness/angularity to both ‘basic’ taste names and actual tastants. In Experiment 1, crossmodal correspondences between taste names and abstract shapes were assessed. Next, the results were replicated in a larger online study (Experiment 2). Experiment 3 assessed the role of liking in the association between taste words and morphed shapes along the roundness/angularity dimension. In Experiment 4, basic tastants were mapped to the roundness/angularity dimension, while the mediating role of liking for each taste was assessed. Across the 4 experiments, participants consistently matched sweetness to roundness. What is more, people’s liking for a taste (but not their liking for imagined tastes) appeared to influence their shape matching responses. These results are discussed in terms of crossmodal correspondences, and a potential role for hedonics is outlined.     

Analysis and design of smart mandrels using shape memory polymers

Based on the thermomechanical mechanism of shape memory polymers (SMPs), the three-dimensional thermomechanical constitutive equation that can be used in the ABAQUS finite element simulation was derived. Then this paper compiled UMAT subroutine and simulated the thermomechanical behaviors of SMP smart mandrels. In addition, the properties of shape fixity and shape recovery ratio of SMP were considered in detail. Finally, filament winding experiments were proceeded on bottle-shaped and air duct-shaped mandrels and the simple and efficient demoulding of SMP mandrels were verified. The results showed the feasibility of SMP as the smart mandrels from practical application in the future.