Fast computation of computer-generated hologram using Xeon Phi coprocessor

We report fast computation of computer-generated holograms (CGHs) using Xeon Phi coprocessors, which have massively x86-based processors on one chip, recently released by Intel. CGHs can generate arbitrary light wavefronts, and therefore, are promising technology for many applications: for example, three-dimensional displays, diffractive optical elements, and the generation of arbitrary beams. CGHs incur enormous computational cost. In this paper, we describe the implementations of several CGH generating algorithms on the Xeon Phi, and the comparisons in terms of the performance and the ease of programming between the Xeon Phi, a CPU and graphics processing unit (GPU).     

Preparation of highly crystalline nanofibers on Fe and Fe-Ni catalysts with a variety of graphene plane alignments

The present study confirmed that highly crystalline nanofibers with controlled structure may be prepared over Fe and Fe-Ni alloy catalysts. The degree of graphitization of various carbon nanofibers (CNFs) was analyzed by using C(0 0 2) peaks from the XRD profiles. The C(0 0 2) peaks of CNFs over Fe catalyst shifted to higher angle and became narrower as the preparation temperature increased from 560 to 620 °C. Tubular CNFs prepared at temperature higher than 630 °C showed lower 2θ angles compared to those of platelet fibers. CNFs prepared over Fe-Ni catalysts tended to resemble those prepared over Fe catalysts. The degree of graphitization of platelet CNFs resembled natural graphite, while d_0 0 2 of the tubular CNFs showed values below the 3.39 Å reported as a theoretical minimum for a cylindrical alignment. Lc_0 0 2 of platelet and tubular CNFs increased by heat treatment at 2000 and 2800 °C though d_0 0 2 changed little. A transverse section of platelet and tubular CNFs had a hexagonal shape, not a round shape. The hexagonal column allows AB stacking of hexagonal planes that can give perfect hexagonal alignment.     

Fabrication of large area diffraction grating using LIGA process

X-ray imaging is a very important technology in the fields of medical, biological, inspection, material science, etc. However, it is not enough to get the clear X-ray imaging with low absorbance. We have produced a diffraction gratings for obtaining high resolution X-ray phase imaging, such as X-ray Talbot interferometer. In this X-ray Talbot interferometer, diffraction gratings were required to have a fine, high accuracy, high aspect ratio structure. Then, we succeeded to fabricate a high aspect ratio diffraction grating with a pitch of 8 μm and small area using a deep X-ray lithography technique. We discuss that the diffraction gratings having a narrow pitch and an large effective area to obtain imaging size of practical use in medical application. If the pitch of diffraction gratings were narrow, it is expected high resolution imaging for X-ray Talbot interferometer. We succeeded and fabricated the diffraction grating with pitch of 5.3 μm, Au height of 28 μm and an effective area of 60 × 60 mm².     

Closed-form forward kinematics solutions of a 4-DOF parallel robot

It is well known that the forward kinematics of parallel robots is a very difficult problem. Closed-form forward kinematics solutions have been reported only to a few special classes of parallel robots. This paper presents closed-form forward kinematics solutions of a 4-DOF parallel robot H4. A 16th order polynomial in a single variable is derived to solve the forward kinematics of the H4. The 16 roots of the polynomial lead to at most 16 different forward kinematics solutions. A numerical verification is also presented.     

Stabilization of geometrically nonlinear topology optimization by the Levenberg–Marquardt method

This paper deals with the design of compliant mechanisms in a continuum-based finite-element representation. Because the displacements of mechanisms are intrinsically large, the geometric nonlinearity is essential for designing such mechanisms. However, the consideration of the geometric nonlinearity may cause some instability in topology optimization. The problem is in the analysis part but not in the optimization part. To alleviate the analysis problem and eventually stabilize the optimization process, this paper proposes to apply the Levenberg–Marquardt method to the nonlinear analysis of compliant mechanisms.     

Clinicopathological significance of intratubular giant macrophages in progressive glomerulonephritis

Very large macrophages, which we have termed "giant macrophages" (G-M phi), have been found in renal tubules, some containing cytoplasmic vacuoles. To elucidate their pathophysiological roles, we examined renal biopsy tissues from various primary glomerulonephritis (GN) and tubulointerstitial nephritis (TIN) using immunohistochemistry with monoclonal antibodies against M phi and other cell surface markers. Giant macrophages were absent or rare in TIN, minimal change nephrotic syndrome, and minor glomerular abnormalities, but G-M phi was plentiful in progressive glomerulonephrides such as IgA nephropathy with crescents, membranoproliferative GN, focal segmental glomerulosclerosis, and especially in crescentic GN. These G-M phi were usually seen in the lumen of renal tubules, but occasionally were found in the Bowman's spaces and glomerular tufts, and similar cells were also found in urine. Moreover, they frequently made contact with tubular epithelial cells expressing intercellular adhesion molecule-1, and the tubular epithelial cells in such lesions often had degenerative changes. Giant M phi may damage tubular epithelial cells from the luminal side. Phenotypically, G-M phi showed activated (CD71+) and mature (25F9+) characteristics along with features of M phi (CD68+), and the cytoplasm contained a great deal of lipids. The numbers of G-M phi in renal tissues closely correlated with the degree of hematuria (rho = 0.5, P < 0.001), serum creatinine value (r = 0.63, P < 0.001) in GN patients (N = 96) and with proteinuria in IgA nephropathy patients (r = 0.89, P < 0.001, N = 27). These data suggest that G-M phi are M phi that were activated and matured in certain active inflammatory sites, which flowed into tubules and then into urine. Thus, the existence of G-M phi in biopsy tissue or urine reflect the activity of GN and may have a predictive value for the progression of GN.     

In situ measurements of ion-exchange processes in single polymer particles: laser trapping microspectroscopy and confocal fluorescence microspectroscopy

Ion-exchange processes of a cationic dye (Rhodamine B; RhB) were studied for individual polymer particles (diameter of 16-20 μm) by laser trapping microspectroscopy and confocal fluorescence laser microspectroscopy. The absorbance of RhB at 565 nm adsorbed on a cation-exchange particle increased linearly with the concentration of RhB in the aqueous phase, while it was independent of the particle diameter. Fluorescence intensity profile measurements of RhB along the particle diameter by confocal fluorescence microspectroscopy directly proved that ion exchange took place in the surface layer (～2-μm thickness) of the particle in the initial stage (1 h). Diffusion of RhB in the particle was very slow, and ion exchange proceeded gradually to the inner volume in the order of days. The ion-exchange processes were analyzed on the basis of simulation of the time course of the concentration profile of RhB in the particle, and the diffusion coefficient of RhB was determined to be (2-4) × 10(-11) cm(2)·s(-1).     

Local normalized linear summation kernel for fast and robust recognition

Kernel-based methods are effective for object detection and recognition. However, the computational cost when using kernel functions is high, except when using linear kernels. To realize fast and robust recognition, we apply normalized linear kernels to local regions of a recognition target, and the kernel outputs are integrated by summation. This kernel is referred to as a local normalized linear summation kernel. Here, we show that kernel-based methods that employ local normalized linear summation kernels can be computed by a linear kernel of local normalized features. Thus, the computational cost of the kernel is nearly the same as that of a linear kernel and much lower than that of radial basis function (RBF) and polynomial kernels. The effectiveness of the proposed method is evaluated in face detection and recognition problems, and we confirm that our kernel provides higher accuracy with lower computational cost than RBF and polynomial kernels. In addition, our kernel is also robust to partial occlusion and shadows on faces since it is based on the summation of local kernels.     

A study of Q-learning considering negative rewards

In the reinforcement learning system, the agent obtains a positive reward, such as 1, when it achieves its goal. Positive rewards are propagated around the goal area, and the agent gradually succeeds in reaching its goal. If you want to avoid certain situations, such as dangerous places or poison, you might want to give a negative reward to the agent. However, in conventional Q-learning, negative rewards are not propagated in more than one state. In this article, we propose a new way to propagate negative rewards. This is a very simple and efficient technique for Q-learning. Finally, we show the results of computer simulations and the effectiveness of the proposed method.