Search

All optical and one dimensional scanning surface measurement system is constructed. The system uses the principleof spectral interference and optical Fourier transform, or the principle of joint transform correlator. It can determinethe three dimensional surface of an object in the accuracy of few 1Om.As a first demonstration, path-length difference of a Michelson interferomter is measured. Next, the surface ofan object which has stepped surface is measured. 1. INTRODUCTION Optical coherence tomography (OCT) has been largely used in the field of in vivo shape measurement.1,'? j this method, a reference beam and an object beam of a Michelson interferometer are interfered on a power meter. When a sample object is placed on the object-arm of the interferometer, the depth information of the object is encodedin the temporal interferogram. Hence, from the signal, we can obtain the shape of the sample with the accuracy determined by the coherence length of the light source. Generally, the OCT technique gives us the depth informationof the object, even if the object is a high scattering object. However, the method demands three dimensional (3-D)

Spatio-temporal pulse shaper which controls the spatial and temporal profile of an ultrafast light pulse. The pulse shaper uses the spatio-temporal coupling effect in a pulse shaper. In this paper, the quantitative property is analyzed numerically by using Wigner distribution function. By the analysis, it is confirmed that spatio-temporal output pulse track forms the differentiation of the phase mask.

We constructed a learning optical neural network with variable learning coefficient by fuzzy controlling. The system performs learning with 2D optical means for handling images without scanning and pixeling. By the fuzzy controlling theory, the learning coefficient in back- propagation algorithm is adjusted based on the training error and training time. The effectiveness of the system confirmed by the learning experiments of the recognition of three human faces.© (2000) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

The method of spatial mapping in biology vision field is applied to artificial neural networks for pattern recognition. By the coordinate transform that is called the complex-logarithm mapping and Fourier transform, the input images are transformed into scale- rotation- and shift- invariant patterns, and then fed into a multilayer neural network for learning and recognition. The results of computer simulation and an optical experimental system are described.© (2000) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

A method to control a 3D pulse shape is proposed. The three dimensions are two spatial dimensions and one temporal dimension. We use the new method to control pulses two dimensionally, and we introduce another new optical system which is the dimensionally extended former system with 2D spatial light modulator.

We describe an optical learning three-layer neural network that uses a back-propagation algorithm for human face recognition. Learning is performed by two-dimensional optical means for handling images without scanning and pixeling, because a face recognition problem needs a large-scale network. The two-dimensional input image is treated directly without vectorizing so that a recognition process of an unknown facial image is executed in real-time. We demonstrate training and recognition of three human faces using an experimental system.

We constructed optical systems for learning neural networks with 2D structure sing Selfoc microlens arrays. On the systems, pattern recognitions of typed alphabetic characters which were directly detected with a CCD camera were realized. Liquid crystal devices, an electron-beam addressed spatial light modulator and a Pockels readout optical modulator were used for displaying weight matrices of the neural networks. The weights were renewed according to the random search algorithm or the delta rule with error signals calculated optically. The 2D structure for image processing can be implemented with large scale networks that consist of several thousands input neurons.

We proposed a new renewal method of the weights to realize a back propagation learning algorithm with a large number of iteration times in an optical neural network. The main purpose of this method is to use the limited dynamic range of an optical device widely. Since some weights displaying on the device are quickly saturated in the conventional method, the learning is not completed in many cases. The new technique can suppress saturation of the weights so that learning is completed. The computer simulation and the optical experiment are presented.

Using an electron-beam addressed spatial light modulator (EBSLM), an optical three-layer neural network with the back-propagation learning algorithm is successfully implemented. The weight matrices are memorized and modified on the EBSLM. To realize a large number of learning iterations, we propose a new renewal method of the weights.© (1992) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Experimental and theoretical investigations on the formationof acceptor- acceptor pair emission in GaAsYunosuke Makita, Masahiko Mori, *N*obukazu Ohnishi,* K* suhiro Irie,*Shigeru Shigetomi, and Hideki TanakaElectrotechnical Laboratory1 -1 -4 Umezono, Sakura -mura, Niihari -gun, Ibaraki 305, JAPAN*Institute of Fundamental Analysis Inc. Ltd.3 -24 -3 Yoyogi, Shibuya -ku, Tokyo 151, JAPAN* *Kurume University, Department of Physics635 Mii- machi, Kurume -shi, Fukuoka 830, JAPAN** *Tokai University1117 Kitakaname, Hiratsuka -shi, Kanagawa 259 -12, JAPANAbstractLow temperature photoluminescence (PL) measurements were carried out for C -, Be -, Mg -,Ge -, Zn- and Cd -doped GaAs prepared by molecular beam epitaxy (MBE), liquid phase epitaxy(LPE) and ion implantation (I4). Two new emissions, temporarily denoted by 'g' and [g -g]were commonly observed in the above materials and no principal difference of spectra was ob-tained among the samples prepared by different doping methods. 'g' is situated just belowthe well -known bound exciton emissions (B.E.) and seems to be unshifted against the increaseof acceptor concentration, [A] and was suggested to be the excited energy state of an iso-lated acceptor. [g -g] presents a significant energy -shift towards lower energy sides withincreasing [A] and at the highest limit of [A], it converges at the well- defined opticaltransition from conduction band to acceptor level, (e,A). These observations together with apreliminary theory suggest that [g -g] can be ascribed to the pair between excited -acceptors.It was for the first time clearly demonstrated that the absence and overlook of these twospecific emissions pertinent to acceptors in conventionally -prepared samples can be at-tributed to the strong optical compensation effect among acceptors and unintentionally in-troduced residual donors.IntroductionIt has been long believed that optical properties of impurity -doped GaAs, especiallythose of acceptor- and donor -doped ones are well characterized and documented. In Fig.1 -a)we show a generally accepted scheme of optical transition in direct band -gap semiconductgrswhere only one species of donor and /or acceptor impurities are assumed to be existing.

Recently we found two new emissions denoted by 'g' and [g-g] in the low temperature photoluminescence (PL) spectra of acceptor-impurity incorporated GaAs. 'g is situated at an energy slightly below that of the bound exciton emissions. [g-g] is situated just below 'g' and shows a significant energy-shift towards the lower energy side with increasing acceptor concentration,[A]. Previously we proposed a model in which [g-g] was ascribed to the acceptor-acceptor pair formed by the overlapping of wave functions of the 2p state of the isolated acceptors. Although the calculated energy as a function of [A] showed qualitative agreement with the experimental results, it was always larger than the observed binding energy of [g-g]. In this paper it is indicated that the red shift of [g-g] with increasing [A], and its energy locking at a critical [A], can be well explained by phenomenologically taking into account the screening effect of the hole which is in the ground state. It was explicitly demonstrated that [g-g] is a very useful optical tool for the estimation of [A].