TCC - Engenharia Eletrônica (UACSA)
URI permanente para esta coleçãohttps://arandu.ufrpe.br/handle/123456789/2917
Navegar
2 resultados
Resultados da Pesquisa
Item Implementação em VHDL e análise comparativa da Transformada Discreta do Cosseno e a aproximação de Lengwehasatit-Ortega(2024-08-27) Silva, Leonardo Nogueira Lindolfo da; Ferreira, Felipe Alberto Barbosa Simão; http://lattes.cnpq.br/9939255113143786; http://lattes.cnpq.br/3309281833356970The high volume of images transmitted daily over the internet consumes a significant portion of bandwidth and processing capacity. To optimize compression and maintain quality, digital signal processing tools like the Discrete Cosine Transform (DCT) are utilized. The DCT is a mathematical operation that concentrates most of the signal's energy in low frequencies, making it highly useful in image compression algorithms. The calculation of the Discrete Cosine Transform is performed using matrix multiplication, where the elements of the transformation matrix are floating-point numbers. To simplify these calculations, various approximations for the Discrete Cosine Transform that use as many integers as possible in their transformation matrices can be found in the literature. Due to the recurrent matrix calculations in modern computational systems, especially for image processing and artificial intelligence, many systems feature ASICs (Application Specific Integrated Circuits) or parts of SoCs (System on Chip) dedicated to this task. In this work, several DCT approximations were evaluated in the context of image compression. The Lengwehasatit-Ortega approximation, which showed the best performance, along with the exact DCT, were implemented in VHDL and synthesized in FPGA. It was observed that, following the same design philosophy, the approximation consumed significantly fewer hardware resources, as expected.Item Filtragem de sinais sobre grafos aplicada à classificação de dados ruidosos definidos em estruturas irregulares(2023-10-03) Medeiros, Jorge Barros; Ferreira, Felipe Alberto Barbosa Simão; http://lattes.cnpq.br/9939255113143786; http://lattes.cnpq.br/9517722947492097Graph convolutional networks (GCN) allow deep neural networks to learn from data defined in irregular structures represented by a graph. This approach has attracted more and more attention in recent years. Such algorithms generally learn based on the information associated with the vertices and edges present in the graph, being possible to use them to improve the performance of several tasks such as classification of vertices, a signal defined on the graph or the graph itself. The objective of this work is to evaluate the classification of noisy data arranged in an irregular structure from the training of low-pass filters performed by the convolutional layers of a GNN. Filters are designed and applied based on signal processing tools on graphs such as the graph Fourier transform. The graph-based convolutional neural network will perform the learning of parameters for different noise scales in the graph Fourier transform domain that are able to extract relevant information for classification of the noisy dataset.