Visual pattern recognition network: Its training algorithm and its optoelectronic architecture

A visual pattern recognition network and its training algorithm are proposed. The network constructed of a one-layer morphology network and a two-layer modified Hamming net. This visual network can implement invariant pattern recognition with respect to image translation and size projection. After supervised learning takes place, the visual network extracts image features and classifies patterns much the same as living beings do. Moreover we set up its optoelectronic architecture for real-time pattern recognition. (C) 1996 Optical Society of America

Prescription for optimizing holograms in LiNbO3 : Fe : Mn

The photorefractive holographic dynamics of grating formation in photochromic doubly doped LiNbO3:Fe:Mn crystal is studied numerically and analytically in terms of the two-center model of Kukhtarev Et al. [Ferroelectrics 22, 949 (1979)]. The relations among the recorded and fixed space-charge fields and the doping densities, the oxidation-reduction states of the fields, and the intensities of UV-sensitizing and red recording beams are studied. Important conditions and effects are feued, and an optimal prescription for material doping and oxidation-reduction processing is suggested in which the crystal can be strongly oxidized and the Mn-doping density is smaller than the Fe-doping density. (C) 2000 Optical Society of America. OCIS codes: 050.7330, 190.5330, 090.2900.

Pentylenetetrazole-induced status epilepticus following training does not impair expression of morphine-induced conditioned place preference

Learning and memory play an important role in morphine addiction. Status epilepticus (SE) can impair the spatial and emotional learning and memory. However, little is known about the effects of SE on morphine-induced conditioned place preference (CPP). Th

Impact of Brunswick river mouth training walls on adjacent beaches, Brunswick heads, New South Wales, Australia

River training walls have been built at scores of locations along the NSW coast and their impacts on shoreline change are still not fully understood. In this study, the Brunswick River entrance and adjacent beaches are selected for examination of the impact of the construction of major training walls. Thirteen sets of aerial photographs taken between 1947 and 1994 are used in a CIS approach to accurately determine tire shoreline Position, beach contours and sand volumes, and their changes in both time and space, and then to assess the contribution of both tire structures and natural hydrodynamic conditions to large scale (years-decades and kilometres) beach changes. The impact of the training walls can be divided into four stages: natural conditions prior to their construction (pre 1959), major downdrift erosion and updrift accretion during and. following the construction of the walls in 1959 similar to 1962 and 1966. diminishing impact of the walls between 1966 and 1987, and finally no apparent impact between 1987 similar to 1994. The impact extends horizontally about 8 km updrift and 17 km downdrift, and temporally up to 25 years..

New method for analysis of Chinese herbal complex prescription and its application

Traditional Chinese medicine (TCM) is a great treasure of China, the analysis of which is an arduous task. The viewpoint that all chemical constituents of Chinese herbal complex prescription should be analyzed as a black box is elucidated for the first time. Intelligent multi-mode multi-column chromatographic system (IMMCC) with its hybrids is the basic method and HPLC Unified Method is the breakthrough for the black box analysis. Dang-Gui-Bu-Xue-Tang was selected as a typical TCM and a systematic separation method from non-aqueous mobile phase to pure water mobile phase was put forward in order to convert unknown sample to known sample. The a, c values and UV spectra of 66 components of Astragalus, 78 components of Angelica and 71 components of Dang-Gui-Bu-Xue-Tang were obtained. Intelligent optimization and peak identification method and software for complex samples were developed and the optimum multi-step multi-binary gradient curve of mobile phase for Astragalus was ascertained. The maximum error and minimum error of predicted retention time for all components of Astragalus are 8.62% and 0.05% respectively. All components of Astragalus were compared with those of Angelica and it is found that many components of Astragalus are the same as those of Angelica, while the contents of these components are different. Many components of Dang-Gui-Bu-Xue-Tang are also the same as those of Astragalus and Angelica with different contents.

Separation of compounds interacting with liposome membrane in combined prescription of traditional Chinese medicines with immobilized liposome chromatography

Immobilized liposome chromatography (ILC), the stationary phase of which has been regarded as a mimic biomembranes system was used to separate and analyze compounds interacting with liposome membrane in Danggui Buxue decoction, a combined prescription of traditional Chinese medicines (CPTCMs), and its compositions Radix Astragli and Radix Angelica Sinensis. More than 10 main peaks in the extract of Danggui Buxue decoction were resolved on the ILC column, suggesting that more than 10 components in the prescription have significant retention on ILC column. Ligustilide, astragaloside, TV and formononetin, three main bioactive ingredients in Danggui Buxue decoction, were found to have relatively significant, while ferulic acid, another bioactive ingredient in the prescription, relatively weak retention on ILC column. Effects of the eluent pH and amount of immobilized phosphatidylcholine (PC) on separation of interactional compounds in the extract of Danggui Buxue decoction were also investigated. It was found that these two factors strongly affected the retention of some interactional compounds. In addition, the fractions partitioned with different solvents from water extract of this combined prescription were evaluated with this ILC column system. (c) 2005 Elsevier B.V. All rights reserved.