An Error Estimation Framework for Many‐Light Rendering

The popularity of many‐light rendering, which converts complex global illumination computations into a simple sum of the illumination from virtual point lights (VPLs), for predictive rendering has increased in recent years. A huge number of VPLs are usually required for predictive rendering at the cost of extensive computational time. While previous methods can achieve significant speedup by clustering VPLs, none of these previous methods can estimate the total errors due to clustering. This drawback imposes on users tedious trial and error processes to obtain rendered images with reliable accuracy. In this paper, we propose an error estimation framework for many‐light rendering. Our method transforms VPL clustering into stratified sampling combined with confidence intervals, which enables the user to estimate the error due to clustering without the costly computing required to sum the illumination from all the VPLs. Our estimation framework is capable of handling arbitrary BRDFs and is accelerated by using visibility caching, both of which make our method more practical. The experimental results demonstrate that our method can estimate the error much more accurately than the previous clustering method.     

Scanning-laser-beam semiconductor pH-imaging sensor

We have constructed a two-dimensional (2-D) pH-imaging sensor that enables us to observe th H⁺ distribution produced by living cells. The pH-sensing principle is similar to that of thelight-addressable potentiometric sensor. The electrolyte-insulator-semiconductor (EIS) structure (electrolyte---Si₃N₄---SiO₂---Si) is illuminated by a focused (1 μm) and modulated (1–10 kHz) He—Ne laser bean from the backside of the semiconductor, and the a.c. photocurrent flowing through the EIS structure is measured. By scanning the laser beam, 2-D pH images can be obtained. By using this sensor, pH distributions of colonies of yeast (Saccharomyces cerevisiae) have been observed. The spatial resolution of this sensor could be improved by thinning the Si wafer.     

Calculating accumulations

Theaccumulation strategy consists of generalizing a function over an algebraic data structure by inclusion of an extra parameter, anaccumulating parameter, for reusing and propagating intermediate results. However, there remain two major difficulties in this accumulation strategy. One is to determinewhere andwhen to generalize the original function. The other, surprisingly not yet receiving its worthy consideration, is how to manipulate accumulations. To overcome these difficulties, we propose to formulate accumulations ashigher order catamorphisms, and provide several general transformation rules for calculating accumulations (i.e., finding and manipulating accumulations) bycalculation-based (rather than a search-based) program transformation methods. Some examples are given for illustration. Zhenjiang Hu, Dr.Eng.: He is an Assistant Professor in Information Engineering at the University of Tokyo. He received his BS and MS in Computer Science from Shanghai Jiao Tong University in 1988 and 1990 respectively, and his Dr. Eng. degree in Information Engineering from the University of Tokyo in 1996. His current research concerns programming languages, functional programming, program transformation, and parallel processing. Hideya Iwasaki, Dr.Eng.: He is an Associate Professor in Information Engineering at the University of Tokyo. He received the M.E. degree in 1985, the Dr. Eng. degree in 1988 from the University of Tokyo. His research interests are list processing languages, functional languages, parallel processing, and constructive algorithmics. Masato Takeichi, Dr.Eng.: He is Professor in Mathematical Engineering and Information Engineering at the University of Tokyo since 1993. After graduation from the University of Tokyo, he joined the faculty at the University of Electro-Communications in Tokyo before he went back to work at the University of Tokyo in 1987. His research concerns the design and implementation of functional programming languages, and calculational program transformation systems.     

Development of an epileptic seizure prediction algorithm using R–R intervals with self-attentive autoencoder

Epilepsy is a neurological disorder that may affect the autonomic nervous system (ANS) from 15 to 20 min before seizure onset, and disturbances of ANS affect R–R intervals (RRI) on an electrocardiogram (ECG). This study aims to develop a machine learning algorithm for predicting focal epileptic seizures by monitoring R–R interval (RRI) data in real time. The developed algorithm adopts a self-attentive autoencoder (SA-AE), which is a neural network for time-series data. The results of applying the developed seizure prediction algorithm to clinical data demonstrated that it functioned well in most patients; however, false positives (FPs) occurred in specific participants. In a future work, we will investigate the causes of FPs and optimize the developing seizure prediction algorithm to further improve performance using newly added clinical data.

     

A recursive construction algorithm for covariance control

This paper proposes an algorithm to compute solutions X to the linear matrix equation and inequality of the type (I-BB⁺)(AX+XA'+W)(I-BB⁺)=0, X>0. This problem arises in the synthesis of covariance controllers; the set of symmetric matrices X assignable as a closed-loop state covariance by a stabilizing controller is characterized by these conditions. Our algorithm generates analytical solutions to the above problem in a recursive manner. In this sense, our algorithm is essentially different from other computational methods pertinent to this problem, such as convex programming. As a result, the algorithm does not involve the issue of convergence and terminates in an a priori known finite number of steps. Thus, the computational complexity is expected to be much less than that of other methods     

Fast Particle‐based Visual Simulation of Ice Melting

The visual simulation of natural phenomena has been widely studied. Although several methods have been proposed to simulate melting, the flows of meltwater drops on the surfaces of objects are not taken into account. In this paper, we propose a particle‐based method for the simulation of the melting and freezing of ice objects and the interactions between ice and fluids. To simulate the flow of meltwater on ice and the formation of water droplets, a simple interfacial tension is proposed, which can be easily incorporated into common particle‐based simulation methods such as Smoothed Particle Hydrodynamics. The computations of heat transfer, the phase transition between ice and water, the interactions between ice and fluids, and the separation of ice due to melting are further accelerated by implementing our method using CUDA. We demonstrate our simulation and rendering method for depicting melting ice at interactive frame‐rates.     

A Modeling and Rendering Method for Snow by Using Metaballs

The display of natural scenes such as mountains, trees, the earth as viewed from space, the sea, and waves have been attempted. Here a method to realistically display snow is proposed. In order to achieve this, two important elements have to be considered, namely the shape and shading model of snow, based on the physical phenomenon. In this paper, a method for displaying snow fallen onto objects, including curved surfaces and snow scattered by objects, such as skis, is proposed. Snow should be treated as particles with a density distribution since it consists of water particles, ice particles, and air molecules. In order to express the material property of snow, the phase functions of the particles must be taken into account, and it is well-known that the color of snow is white because of the multiple scattering of light. This paper describes a calculation method for light scattering due to snow particles taking into account both multiple scattering and sky light, and the modeling of snow.     

Toward Optimal Space Partitioning for Unbiased,Adaptive Free Path Sampling of Inhomogeneous Participating Media

Photo‐realistic rendering of inhomogeneous participating media with light scattering in consideration is important in computer graphics, and is typically computed using Monte Carlo based methods. The key technique in such methods is the free path sampling, which is used for determining the distance (free path) between successive scattering events. Recently, it has been shown that efficient and unbiased free path sampling methods can be constructed based on Woodcock tracking. The key concept for improving the efficiency is to utilize space partitioning (e.g., kd‐tree or uniform grid), and a better space partitioning scheme is important for better sampling efficiency. Thus, an estimation framework for investigating the gain in sampling efficiency is important for determining how to partition the space. However, currently, there is no estimation framework that works in 3D space. In this paper, we propose a new estimation framework to overcome this problem. Using our framework, we can analytically estimate the sampling efficiency for any typical partitioned space. Conversely, we can also use this estimation framework for determining the optimal space partitioning. As an application, we show that new space partitioning schemes can be constructed using our estimation framework. Moreover, we show that the differences in the performances using different schemes can be predicted fairly well using our estimation framework.     

Fracture toughness of sea ice

Linear elastic fracture mechanics, which has been applied to fragile substances and successfully used for studying the brittle fracture of metallic materials, was utilized to gain an understanding of the fracture phenomena of sea ice.The present paper reports the first results of investigations into the fracture-toughness value of sea ice, which was analyzed experimentally as a function of strain rate on the basis of the stress-intensity-factor concept.The fracture toughness, K_IC, of sea ice, which was measured by an in-situ three-point bending test on notched specimens, shows almost constant value if the strain rate is less then 10^?3 s^?1 and decreases with increasing strain rate if the strain rate exceeds 10^?3 s^?1. K_IC data show considerably less scatter than existing data such as the compressive, tensile and flexural strengths.It was confirmed in the present study that the linear elastic-fracture-mechanics concept is effective for analyzing the fracture phenomena of sea ice. Moreover, the K_IC value was shown to be closely related to sea-ice structures (e.g. the size of crack-like flaws such as brine cells).It is also suggested that the fracture-toughness test might prove to be a standard testing method to obtain the sea-ice strength, since once K_1C and crack-like flaw sizes are determined, the less-scattered critical-fracture stress can be calculated.