用金字塔图法匹配边界图象的三眼立体视觉方法

本文在分析了双眼立体视觉面临的本质困难之后，提出了用金字塔图法匹配特征基元的三眼立体视觉方法，这种方法由于建筑在透视投影几何原理的稳固基础上，无须任何先验限制，也不以图象相似性为基础，故可较好地解决精度与匹配难度的矛盾，且运用金字塔图法进行匹配传播使在保证精度的基础上，匹配速度大大提高，对计算机人工图象的匹配结果表明，误匹配率几乎为零，且有较高的匹配速度和精度。     

Utilizing an Animal Model to Identify Brain Neurodegeneration-Related Biomarkers in Aging

Glycine N-methyltransferase (GNMT) regulates S-adenosylmethionine (SAMe), a methyl donor in methylation. Over-expressed SAMe may cause neurogenic capacity reduction and memory impairment. GNMT knockout mice (GNMT-KO) was applied as an experimental model to evaluate its effect on neurons. In this study, proteins from brain tissues were studied using proteomic approaches, Haemotoxylin and Eosin staining, immunohistochemistry, Western blotting, and ingenuity pathway analysis. The expression of Receptor-interacting protein 1(RIPK1) and Caspase 3 were up-regulated and activity-dependent neuroprotective protein (ADNP) was down-regulated in GNMT-KO mice regardless of the age. Besides, proteins related to neuropathology, such as excitatory amino acid transporter 2, calcium/calmodulin-dependent protein kinase type II subunit alpha, and Cu-Zn superoxide dismutase were found only in the group of aged wild-type mice; 4-aminobutyrate amino transferase, limbic system-associated membrane protein, sodium- and chloride-dependent GABA transporter 3 and ProSAAS were found only in the group of young GNMT-KO mice and are related to function of neurons; serum albumin and Rho GDP dissociation inhibitor 1 were found only in the group of aged GNMT-KO mice and are connected to neurodegenerative disorders. With proteomic analyses, a pathway involving Gonadotropin-releasing hormone (GnRH) signal was found to be associated with aging. The GnRH pathway could provide additional information on the mechanism of aging and non-aging related neurodegeneration, and these protein markers may be served in developing future therapeutic treatments to ameliorate aging and prevent diseases.     

Partition model into 3D-printable and no supporting parts

The printing object must fit into the printing working volume and overhangs require a disposable support structure to be added, which are two main problems in the 3D printing process. Existing algorithms cannot solve these two problems at the same time. To solve these problems, we present a model partition algorithm, dividing the model into the pyramidal fitting printing working volume, with the pyramidal having the shape which can be printed without a supporting structure. Firstly, we partition the model surface using the region growing method and analyze the region's normal vector to determine the candidate dividing directions. Secondly, we use the candidate dividing directions to generate candidate dividing planes in order to segment the model. If the divided sub-model is not a pyramidal,continue segmenting the sub-model by using the same method until all of the sub-models are pyramidal. The candidate dividing planes may generate multi-group division modes. Each division mode constructs a tree, the evaluation function is employed to appraise the dividing values and the beam search method is utilized to search the largest value tree in the solution space which is the optimal partition. Experimental results show that the proposed algorithm can divide the model into sub-models which needn't support structures and fit into the printing working volume.     

An asynchronous recurrent linear threshold network approach to solving the traveling salesman problem

In this paper, an approach to solving the classical Traveling Salesman Problem (TSP) using a recurrent network of linear threshold (LT) neurons is proposed. It maps the classical TSP onto a single-layered recurrent neural network by embedding the constraints of the problem directly into the dynamics of the network. The proposed method differs from the classical Hopfield network in the update of state dynamics as well as the use of network activation function. Furthermore, parameter settings for the proposed network are obtained using a genetic algorithm, which ensure a stable convergence of the network for different problems. Simulation results illustrate that the proposed network performs better than the classical Hopfield network for optimization.     

一种基于脉冲耦合神经网络的语音情感识别新方法

针对汉语语音情感识别问题,提出了一种基于脉冲耦合神经网络(PCNN)的识别方法。该方法将语音转化为语谱图后输入到PCNN,得到输出图像的神经元点火序列及其熵序列作为语音情感的特征,利用其特征实现语音情感识别。实验结果表明,该方法可以有效地识别“高兴”与“平常”这两种不同的情感。该方法将PCNN引入到语音情感识别的应用研究中,开拓了语音和图像信号结合处理的新领域,同时对于PCNN的理论研究和实际应用具有重要的现实意义。     

NMR microscopy of single neurons at 500 MHZ

Previous NMR microimaging studies at 360 MHz have demonstrated a clear differentiation between the nucleus and cytoplasm in isolated single neurons. In particular, theT ₂ of the cell nucleus is 2.5 times larger than that of the cytoplasm. In order to determine the magnitude of possibleT ₂ ^* influences on these observations, images of single cells have been obtained at 500 MHz using spin-echo and line-narrowing sequences. Comparison of the images acquired by the two sequences, and of the spin-echo images at 360 and 500 MHz, imply that anyT ₂ ^* contributions are relatively small. Consequently, the measuredT ₂ differences in spin-echo imaging represent a true difference in theT ₂ relaxation in the two cellular compartments.     

应用Simulink模拟耦合神经元的同步活动

癫痫是一组由神经元异常放电所引起的短暂中枢神经系统功能异常为特征的慢性脑部疾病。其发病的原因正是由于神经元集体的异常同步放电行为所致。以FitzHugh-Nagumo神经元为例,使用simulink对耦合神经元的同步活动进行模拟,供相关研究人员进行研究。     

基于代数神经网络的不确定数据知识获取方法

定义了代数神经元、代数神经网络,讨论了不确定数据知识获取的数学机理,设计出一类单输入,单输出的三层前向网络来获取知识的代数神经网络模型,给出一种基于代数神经网络知识获取的方法,通过该网络的学习,能确定任意一组给定数据的目标函数的逼近式。     

Contributions of distinct prefrontal neuron classes in reward processing

The prefrontal cortex(PFC)is thought to be involved in higher order cognitive functions,such as in working memory,abstract categorization,and reward processing.It has been reported that two distinct neuron classes(putative pyramidal cells and interneurons)in the PFC played different functional roles in neural circuits involved in forming working memory and abstract categories.However,it remains elusive how the two types of neurons process reward information in the PFC.To investigate this issue,the activity of single neurons was extracellularly recorded in the PFC of the monkey performing a reward predicting task.PFC neurons were classified into putative pyramidal cells and interneurons,respectively,based on the waveforms of action potentials.Both the two types of neurons encoded reward information and discriminated two reward conditions(the preferred reward condition vs.the nonpreferred reward condition).However,the putative pyramidal neurons had better and more reliable discriminability than the putative interneurons.Also,the pyramidal cells represented reward information in the preferred reward condition,but not in the nonpreferred reward condition by raising their firing rates relative to the baseline rates.In contrast,the interneurons encoded reward information in the nonpreferred reward condition,but not in the preferred reward condition by inhibiting their discharge rates relative to the baseline rates.These results suggested that the putative pyramidal cells and interneurons had complementary functions in reward processing.These findings may help to clarify individual functions of each type of neurons in PFC neuronal circuits involved in reward processing.     

Single neurons in the medial prefrontal cortex of the rat exhibit tonic and phasic coding during trace fear conditioning.

Trace fear conditioning is a learning task that requires the association of an auditory conditioned stimulus (CS) and a shock unconditioned stimulus (US) that are separated by a 20-s trace interval. Single neuron activity was recorded from the prelimbic and infralimbic areas of the medial prefrontal cortex in rats during trace fear conditioning or nonassociative unpaired training. Prelimbic neurons showed learning-related increases in activity to the CS and US, whereas infralimbic neurons showed learning-related decreases in activity to these stimuli. A subset of prelimbic neurons exhibited sustained increases in activity during the trace interval. These sustained prelimbic responses may provide a bridging code that allows for overlapping representations of CS and US information within the trace fear conditioning circuit. (PsycINFO Database Record (c) 2010 APA, all rights reserved)