Short utterance recognition using a network with minimum training

A feedforward network is used to recognize short, digitized, isolated utterances. A high, multispeaker recognition rate is achieved with a small vocabulary with a single training utterance. This approach makes use of the pattern recognition property of the network architecture to classify different temporal patterns in the multidimensional feature space. The network recognizes the utterances without the need of segmentation, phoneme identification, or time alignment. We train the network with four words spoken by one single speaker. The network is then able to recognize 20 tokens spoken by 5 other speakers. We repeat the above training and testing procedure using a different speaker's utterances for training each time. The overall accuracy is 97.5%. We compare this approach to the traditional dynamic programming (DP) approach, and find that DP with slope constraints of 0 and 1 achieve 98.5% and 85% accuracies respectively. Finally we validate out statistics by training and testing the network of a four-word subset of the Texas Instruments (Tl) isolated word database. The accuracy with this vocabulary exceeds 96%. By doubling the size of the training set, the accuracy is raised to 98%. Using a suitable threshold, we are able to raise the accuracy of one network from 87% to 98.5%. Thresholding applied to all networks would then raise the overall accuracy to well over 99%.

This technique is especially promising because of the low overhead and computational requirements, which make it suitable for a low cost, portable, command recognition type of application.     

adaptive fourier threshold filtering: A method to reduce noise and incoherent artifacts in high resolution cardiac images

We introduce an image processing method which reduces white noise and random artifacts in sets of high resolution, time resolved images. At each pixel, the processing consists of: (1) the isolation of a time intensity curve (TIC), (2) Fourier transformation of each TIC, (3) application of a threshold to remove low intensity coefficients, (4) inverse transformation to generate noise reduced TICS which are recombined to form images with improved signal-to-noise ratio (SNR). Noise filtering by Fourier thresholding is demonstrated on a set of cardiac images, resulting in a reduction of the noise energy by approximately 90%.     

Development of phantom for radiographic image optimization of standard patient in the lateral view of chest and skull examination

We present the construction of a homogeneous phantom to be used in simulating the scattering and absorption of X-rays by a standard patient chest and skull when irradiated laterally. This phantom consisted of lucite and aluminium plates with their thickness determined by a tomographic exploratory method applied to the anthropomorphic phantom. Using this phantom, an optimized radiographic technique was established for chest and skull of standard sized patient in lateral view. Images generated with this optimized technique demonstrated improved image quality and reduced radiation doses.     

人脑桥内部的动脉供应和毛细血管构筑

采用血管灌注和透明切片,在立体解剖显微镜下对30例脑桥内部动脉的分群,经行分支和供应范围进行了观察。并应用图像分析系统测算了脑桥内部主要神经核团和纤维束的毛细血管密度。脑桥内部前正中群有三组长动脉,分别由脚间窝下端,基底沟中部和盲孔处穿入,这些动脉在向脑桥被盖部经行和分布的过程中,可形成多级直角分支,使之呈“垂柳状”外观。脑桥内部的毛细血管构筑,在透明切片上,可显示出神经结构的轮廓。其中上橄榄核的毛细血管密度最高,小脑上脚的毛细血管密度最低。     

不同形状界面对界面射流的影响──彩色多普勒与激光荧光染料方法的研究

本文通过应用彩色多普勒及激光荧光染料显示射流方法研究了不同形状界面对射流的影响。结果表明对照组及平坦界面射流时，彩色多普勒与激光荧光染料两种方法无显著差异；在弯曲界面条件下，彩色多普勒射流小于激光荧光染料射流（Ｐ＜０．０１～０．０００１）。彩色多普勒研究表明不同形状界面射流大小分别与每搏容量相关好（ｒ＝０．８３～０．９９），但射流大小明显不同。凹形界面射流面积最小，平坦及凸形界面射流面积也小于对照组。研究提示不同形状界面可以导致界面射流不同程度地减小，并且彩色多普勒也不易对弯曲界面的射流完全成像。在临床上应用彩色多普勒对返流定量时，需考虑上述影响。     

Flow karyotypes and chromosomal DNA contents of genus Triticum species and rye (Secale cereale)

The flow cytometry and chromosome imaging method were jointly used for analyzing genome content and chromosomal DNA content of hexaploid wheat (AABBDD), hexaploid triticale (AABBRR), tetraploid wheat (AABB), and AA, BB, DD genome donors and RR genome rye. Their genome sizes were 34.4 pg, 40.9 pg, 26.2 pg, 12.1 pg, 13.7 pg, 10.5 pg, and 16.9 pg, respectively. The 2C nuclear DNA content of BB genome donor with 13.7 pg was the highest value among the other genome donors, AA or DD. The genome content of tetraploid wheat, unlike hexaploid wheat or hexaploid triticale, was larger than the sum of the genomes of AA and BB genome donors. The DNA content of each chromosome ranged from 1.22 pg in DD genome donor to 2.61 pg in rye. Each chromosome peak was divided into three to four groups. Only one chromosome was included in the highest chromosomal DNA peak in hexaploid wheat, tetraploid wheat, DD genome donor and rye but two chromosomes in AA, BB genome donors, and hexaploid triticale. Correlation between 2C nuclear DNA content and chromosome density volume was the highest value compared with the other chromosomal parameters of chromosome area, or chromosome length.     

PQ系列CT机的数据采集和图像重建系统原理

本文介绍了数据采集系统、图像重建和显示系统的具体组成及完成的相应功能.     

Bilateral field advantage in visual crowding

Thirty randomly oriented T’s were presented in a circle around fixation at an eccentricity of 11° such that each T was crowded by its neighbors. Two locations within the same hemifield (unilateral condition) or one location in each hemifield (bilateral condition) were precued for subsequent probing. Observers were then asked to report the orientation of a target T at one of these locations. A bilateral field advantage was found: target identification was better when the two precued targets were in different hemifields than when they were within the same hemifield. This bilateral advantage was absent when only targets were presented, without any distracters. Further controls showed that this advantage could not be attributed to differences between horizontal and vertical target alignments or to visual field anisotropies. A similar bilateral advantage has been reported for multiple object tracking (Alvarez, G. A., & Cavanagh, P. (2005). Independent resources for attentional tracking in the left and right visual fields. Psychological Science 16(8), 637-643) and other attentional tasks. Our results suggest that crowding also demonstrates separate attentional resources in the left and right hemifields. There was a cost to attending to two targets presented unilaterally over attending to a single target. However, this cost was reduced when the two crowded targets were in separate hemifields.     

Information Fusing Recognition of Traditional Chinese Medicine （TCM） Pulse State Based on Stochastic Fuzzy Neural Network

Based on the fuzzy characteristic of the pulse state and syndromes differentiation thinking mode of TCM, an information fusing recognition method of pulse states based on SFNN （Stochastic Fuzzy Neural Network） is presented in this paper. With the learning ability in parameters and structure, SFNN fuses the measurement information of three pulse-state sensors distributed in Cun, Guan, and Chi location of body for the pulse state recognition. The experimental results show that the percentage of correct recognition with new method is higher than that by single-data recognition one, with fewer off-line train numbers.