Two‐way multi‐view 2‐D/3‐D display combining LC lens and HVxDP panel using novel pixel arrangement

We have developed a two‐way multi‐view 2‐D/3‐D display combining a liquid crystal lens and horizontally and vertically x times‐density pixels (HVxDP) arrangement. The two‐way multi‐view display features the same display resolution in 2‐D and 3‐D modes and a quite small color moiré for landscape and portrait, respectively, when using the HVxDP arrangement. In this paper, we realized suitable 3‐D properties for achieving a good balance between 3‐D moiré and 3‐D crosstalk for landscape and portrait by a two‐way liquid crystal lens with two kinds of focal lengths for the edge part and the center part of the lens.     

Combination of bottleneck feature extraction and dereverberation for distant-talking speech recognition

The performance of speech recognition in distant-talking environments is severely degraded by the reverberation that can occur in enclosed spaces (e.g., meeting rooms). To mitigate this degradation, dereverberation techniques such as network structure-based denoising autoencoders and multi-step linear prediction are used to improve the recognition accuracy of reverberant speech. Regardless of the reverberant conditions, a novel discriminative bottleneck feature extraction approach has been demonstrated to be effective for speech recognition under a range of conditions. As bottleneck feature extraction is not primarily designed for dereverberation, we are interested in whether it can compensate for other carefully designed dereverberation approaches. In this paper, we propose three schemes covering both front-end processing (cascaded combination and parallel combination) and back-end processing (system combination). Each of these schemes integrates bottleneck feature extraction with dereverberation. The effectiveness of these schemes is evaluated via a series of experiments using the REVERB challenge dataset.     

Parallelized fusion visualization system

The speed-up of supercomputers has increased the complexity of simulations. To analyze such kind of data, we believe that new types of visualization software are needed. Therefore, we have been developing a visualization system called “Fusion Visualization”, and the progresses were reported in the AROB 18th and 19th International Symposiums. We introduced the overall concept at the AROB 18th International Symposium, and then demonstrated a sample of flow visualization in a blood vessel in the AROB 19th International Symposium. To extend our system to enable the handling of larger data, we have implemented the proposed system on a parallelized visualization system; AVS/Express PCE (Parallel Cluster Edition). This paper describes the implementation and the benchmark results.     

Relationship between crystallinity and device characteristics of In‐Sn‐Zn oxide

We have reported that the transistors having the c‐axis‐aligned crystalline (CAAC) In‐Ga‐Zn oxide (IGZO) show good performance. Recently, In‐Sn‐Zn Oxide (ITZO) has attracted much attention because of its high electron mobility, as well as IGZO. However, it has been reported that ITZO field effect transistors (FET) tend to have positive Vth (normally‐on characteristics) and poor reliability compared with IGZO‐FETs. We have reported that high‐performance and high‐reliability OS‐FETs can be fabricated by using CAAC‐IGZO, which has high crystallinity and has no clear grain boundaries, as an active layer. Therefore, we have fabricated CAAC‐ITZO thin films to improve performance of ITZO‐FETs by using CAAC‐ITZO as an active layer. In addition, FETs employing CAAC‐ITZO have better characteristics and reliability than FETs using nano‐crystal ITZO. Furthermore, constant photocurrent method (CPM) measurement was carried out in order to estimate density of deep‐level defect states caused by oxygen vacancies in oxide semiconductors. The results show that CAAC‐ITZO has lower density of deep‐level defect states than nano‐crystal ITZO. We attribute the improvement in reliability of ITZO‐FETs to a decrease in deep‐level defect states of an ITZO active layer, as is the case with IGZO.     

An architecture framework for an adaptive extensible processor

To improve the performance of embedded processors, an effective technique is collapsing critical computation subgraphs as application-specific instruction set extensions and executing them on custom functional units. The problem with this approach is the immense cost and the long times required to design a new processor for each application. As a solution to this issue, we propose an adaptive extensible processor in which custom instructions (CIs) are generated and added after chip-fabrication. To support this feature, custom functional units are replaced by a reconfigurable matrix of functional units (FUs). A systematic quantitative approach is used for determining the appropriate structure of the reconfigurable functional unit (RFU). We also introduce an integrated framework for generating mappable CIs on the RFU. Using this architecture, performance is improved by up to 1.33, with an average improvement of 1.16, compared to a 4-issue in-order RISC processor. By partitioning the configuration memory, detecting similar/subset CIs and merging small CIs, the size of the configuration memory is reduced by 40%.     

Fabrication of large-volume microfluidic chamber embedded in glass using three-dimensional femtosecond laser micromachining

We report on fabrication of large-volume, square-shaped microfluidic chamber embedded in glass by scanning a tightly focused femtosecond laser beam inside a porous glass immersed in water. After the hollow structure is created in the porous glass substrate, the fabricated glass sample is post-annealed at 1,050°C during which it can be sintered into a compact glass. By the use of this technique, a 1 mm × 1 mm × 100 μm microchamber connected to four microfluidic channels is created inside the transparent glass substrate, showing that our technique allows for fabrication of not only thin channel structures with arbitrary lengths and configurations, but also hollow structures with infinitely large sizes.     

Particle-based multiple irregular volume rendering on CUDA

In this paper, we describe an improved particle-based volume rendering (PBVR) technique for previewing a large irregular volume dataset using the CUDA architecture. This technique allows for opaque and emissive particles to render translucent volumes without visibility sorting. Our GPU acceleration of PBVR provides the multi-volume rendering feature while remaining compatible with both regular and irregular volumes. We also reduce the memory cost required for storing all sub-pixel values by proposing a pixel repetition technique for a large sub-pixel level. By adjusting the repetition level, we achieved a very smooth level of detail (LOD) control for trading quality for speed. Our work demonstrates a full-detail rendering rate from 5 to 10 fps for irregular volume data with mega-scale cell numbers on an NVIDIA GeForce 8800GTS.     

Solving the irregular strip packing problem via guided local search for overlap minimization

The irregular strip-packing problem (ISP) requires a given set of non-convex polygons to be placed without overlap within a rectangular container having a fixed width and a variable length, which is to be minimized. As a core sub-problem to solve ISP, we consider an overlap minimization problem (OMP) whose objective is to place all polygons into a container with given width and length so that the total amount of overlap between polygons is made as small as possible. We propose to use directional penetration depths to measure the amount of overlap between a pair of polygons, and present an efficient algorithm to find a position with the minimum overlap for each polygon when it is translated in a specified direction. Based on this, we develop a local search algorithm for OMP that translates a polygon in horizontal and vertical directions alternately. Then we incorporate it in our algorithm for OMP, which is a variant of the guided local search algorithm. Computational results show that our algorithm improves the best-known values of some well-known benchmark instances.     

Optimal linear combination of neural networks for improvingclassification performance

This paper presents a new method for linearly combining multiple neural network classifiers based on the statistical pattern recognition theory. In our approach, several neural networks are first selected based on which works best for each class in terms of minimizing classification errors. Then, they are linearly combined to form an ideal classifier that exploits the strengths of the individual classifiers. In this approach, the minimum classification error criterion is utilized to estimate the optimal linear weights. In this formulation, because the classification decision rule is incorporated into the cost function, a more suitable better combination of weights for the classification objective could be obtained. Experimental results using artificial and real data sets show that the proposed method can construct a better combined classifier that outperforms the best single classifier in terms of overall classification errors for test data