Search

Infrapatellar fat pad (IPFP) can be easily obtained during knee surgery, which avoids the damage to patients for obtaining IPFP. Infrapatellar fat pad adipose-derived stem cells (IPFP-ASCs) are also called infrapatellar fat pad mesenchymal stem cells (IPFP-MSCs) because the morphology of IPFP-ASCs is similar to that of bone marrow mesenchymal stem cells (BM-MSCs). IPFP-ASCs are attracting more and more attention due to their characteristics suitable to regenerative medicine such as strong proliferation and differentiation, anti-inflammation, antiaging, secreting cytokines, multipotential capacity, and 3D culture. IPFP-ASCs can repair articular cartilage and relieve the pain caused by osteoarthritis, so most of IPFP-related review articles focus on osteoarthritis. This article reviews the anatomy and function of IPFP, as well as the discovery, amplification, multipotential capacity, and application of IPFP-ASCs in order to explain why IPFP-ASC is a superior stem cell source in regenerative medicine.

Cellulose/Ag/polyaniline conductive composite with rather excellent electrical conductivity was heterogeneously synthesized in this paper. The UV-Vis analysis indicated that homogeneous nanoAg particles deposited on the surface of cellulose in the form of globe particles. They offered some electrons to polyaniline chains. This behavior resulted to the facts that more polyaniline embedded on cellulose and an integrated electrically conductive network formed. Consequently, the high electrical conductivity of the composite was observed. The value was 3.48 S/cm, which was higher two magnitudes than the electrical conductivity of cellulose/polyaniline composite (2.15×10-2S/cm), and even was higher than the electrical conductivity of pure polyaniline (0.142 S/cm). This paper provided a facile method for the preparation of cellulose/Ag/ polyaniline composite with favorable electrical conductivity.

In this paper we consider two kinds of neural networks in which the activation function of each neuron is a multiple valued, piecewised-constant function. The main advantages of the proposed models are that they can store patterns with different grey levels, and that they can store binary patterns with much fewer neurons than the existing models. We prove theoretically the convergence property of the proposed models. Different synthesis methods are developed to guarantee the storage of desired patterns as asymptotic equilibria. Simulation results confirm the effectiveness of the new models.

System behaviors are implemented through interacting between components during the method of component-based software development. Therefore, how to formal the component behaviors have become an important problem. This paper presents a component behavioral modeling method based on event track, which utilize a process with the event track to describe the behavior of component, the event set may be involved in the components were extended to the pointed set, the possible behavior of component in a given environment seen as an event track. This method can effectively solve parallel composition of component behaviors with synchronization constraints. Finally, this paper illustrates concrete details by an example.

We present a new type of Hopfield feedback network for image restoration, called the multiple-valued neural networks (MVFN). The main advantages of this model are that it can store patterns having different grey levels, and that it can store binary patterns with much less neurons than that of a Hopfield binary NN. We apply this new network for noise removal on two different images.

The global convergence of neural networks is known to be the basis of successful applications of neural networks in various computation and recognition tasks. However, almost all the previous studies on neural networks assumed that the interconnection matrix is symmetric. In this paper, we investigate the sufficient condition to guarantee a class of nonlinear continuous neural networks including the Hopfield model as a special case to be global convergent towards unique stable equilibrium point without the assumption of symmetric interconnection. And we also give the sufficient condition to ensure the global convergence of the networks with symmetric interconnection matrix. >

We propose a modified multiple-valued recurrent correlation associative memory with exponential weighting function (MEMRCAM) which is a network model with high storage capacity and strong error correction capability. The operations of this kind of networks depend on the similarity-measure computation and the input-output relation of each neuron being step function. This new model adopts the same principle as that proposed in the previous work (Chiueh and Goodman, 1991). We show the convergent property of the MEMRCAM and simulation result verifies the strong error correction capability with large storage capacity of the MEMRCAM. >

A multiple-valued asymmetric bidirectional associative memory (MABAM) is proposed. Using multiple-valued neurons, this model can store multiple-valued data pairs, so as to decrease the number of neurons in each layer and hence the number of interconnections. We suggest a learning rule for this model and prove its validity. Simulation results show the performance of this model. >

Artificial neural network has been, used as a powerful tool for pattern classification. However, it is difficult to train when the data exhibit non-sparse or overlapping pattern classes which is often the case in practical applications. In this paper, we introduce the feedforward neural network with the hidden layer consisting of multilevel neurons. The convergence property of one-layer neural network with multilevel neurons is proved. The new feedforward model is inherently capable of fuzzy pattern classification of non-sparse or overlapping pattern classes. As an application, we apply the network for the classification of LANDSAT TM data. The results show that this approach produces better results compared with conventional neural networks.