No description available.

No description available.

In our research, we proposed a novel invariant in 2-D image contour recognition based on Hopfield-Tank neural network. At first, we searched the feature points, the position of feature points where are included high curvature and corner on the contour. We used polygonal approximation to describe the image contour. There have two patterns we set, one is model pattern another is test pattern. The Hopfield-Tank network was employed to perform feature matching. In our results show that we can overcome the test pattern which consists of translation, rotation, scaling transformation and no matter single or occlusion pattern.

Submitted by Biblioteca do Instituto de Física (bif@ndc.uff.br) on 2017-07-04T18:39:14Z No. of bitstreams: 1 Dissertaçao_PierreSoares (1).pdf: 729793 bytes, checksum: 12e5898e33b9602f7b327e003b58716b (MD5) / Made available in DSpace on 2017-07-04T18:39:14Z (GMT). No. of bitstreams: 1 Dissertaçao_PierreSoares (1).pdf: 729793 bytes, checksum: 12e5898e33b9602f7b327e003b58716b (MD5) / Conselho Nacional de Desenvolvimento Científico e Tecnológico / Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro / O uso de redes neurais na solução de problemas é bastante atrativa pois suas características possibilitam desempenho superior ao de métodos convencionais [1]. Dentre os diferentes modelos de redes neurais, o modelo de Hopfield apresenta uma grande importância histórica nesse campo. Neste trabalho estudaremos o problema da capacidade de reconhecimento em uma rede de Hopfield utilizando técnicas de análise de tamanho finito. Vamos comparar os resultados obtidos por diferentes métodos com o intuito de obter o valor de [alfa c], o ponto onde a rede passa por uma transição de fase. Para isso utilizaremos simulações computacionais de redes de Hopfield. / The use of neural networks in problem solution is quite attractive because its characteristics enable superior performance than the conventional methods [1]. Among the different models of neural networks, the Hopfield model has a great historic importance in this field. In this work we will study the capacity problem of a Hopfield network by using finite-size analysis. We will compare the results obtained by different methods to find the value of [alpha c], the point where the network undergoes a phase transition. For this we will use computational simulations of Hopfield networks.

Submitted by Biblioteca do Instituto de Física (bif@ndc.uff.br) on 2017-07-14T18:59:48Z No. of bitstreams: 1 Dissertaçao_PierreSoares (1).pdf: 729793 bytes, checksum: 12e5898e33b9602f7b327e003b58716b (MD5) / Made available in DSpace on 2017-07-14T18:59:48Z (GMT). No. of bitstreams: 1 Dissertaçao_PierreSoares (1).pdf: 729793 bytes, checksum: 12e5898e33b9602f7b327e003b58716b (MD5) / Conselho Nacional de Desenvolvimento Científico e Tecnológico / Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro / O uso de redes neurais na solução de problemas é bastante atrativa pois suas características possibilitam desempenho superior ao de métodos convencionais [1]. Dentre os diferentes modelos de redes neurais, o modelo de Hopfield apresenta uma grande importância histórica nesse campo. Neste trabalho estudaremos o problema da capacidade de reconhecimento em uma rede de Hopfield utilizando técnicas de análise de tamanho finito. Vamos comparar os resultados obtidos por diferentes métodos com o intuito de obter o valor de , o ponto onde a rede passa por uma transição de fase. Para isso utilizaremos simulações computacionais de redes de Hopfield. / The use of neural networks in problem solution is quite attractive because its characteristics enable superior performance than the conventional methods [1]. Among the different models of neural networks, the Hopfield model has a great historic importance in this field. In this work we will study the capacity problem of a Hopfield network by using finite-size analysis. We will compare the results obtained by different methods to find the value of , the point where the network undergoes a phase transition. For this we will use computational simulations of Hopfield networks.

There are three main results in this dissertation. They are PLS-completeness of discrete Hopfield network convergence with eight different restrictions, (degree 3, bipartite and degree 3, 8-neighbor mesh, dual of the knight's graph, hypercube, butterfly, cube-connected cycles and shuffle-exchange), exponential convergence behavior of discrete Hopfield network, and simulation of Turing machines by discrete Hopfield Network.

Stuttgart, Univ., Fakultät Elektrotechnik, Diplomarb., 1998.

Techn. University, Diss., 2000--Berlin.

No description available.

No description available.