Ciência da Computação
URI permanente desta comunidadehttps://repositorio.fei.edu.br/handle/FEI/342
Navegar
5 resultados
Resultados da Pesquisa
Trabalho de Conclusão de Curso Reconstrução 3D de ambientes externos baseada na correlação de poucas imagens de diferentes perspectivas(2024-12-11) Medrano, Alessandro Simões; Cunha, Daniel Alves; Mendes, João Vitor Simões; Videira, Pedro Bazaluk MachadoReconstrução 3D está em grande visibilidade atualmente, por conta do avanço tanto da tecnologia de hardwares quanto de softwares. Esse processo tem como propósito trazer novas experiências imersivas através da sua aplicação em diversas áreas, como, por exemplo, na área médica, entretenimento, e construção civil. Porém, há uma limitação na utilização de reconstrução 3D para a criação de malhas de ambientes externos, uma vez que a forma mais comum para a realização deste processo, considerando todas as perspectivas de um ambiente, dá-se por meio da utilização de sensores, os quais realizam o cálculo das profundidades do ambiente, o que facilita sua reconstrução, mas torna-se uma prática restrita por conta do custo dessa tecnologia. Além disso, grande parte das reconstruções 3D de um ambiente externo, a partir de imagens, são realizadas considerando apenas uma perspectiva e com a utilização de uma grande quantidade de dados processados referentes às imagens de um ambiente. Assim, este trabalho propõe a reconstrução 3D de ambientes externos a partir de poucas imagens com a completude de várias perspectivas e sem o uso de sensores. Trata-se de uma metodologia distinta daquelas do estado-da-arte. Os resultados alcançados mostraram que a metodologia proposta aqui é capaz de reconstruir ambientes externos a partir de poucas imagens de diferentes perspectivas, gerando malhas 6D visualmente realistas e em tempo real.- A Bio-Inspired Methodology for Digital Imaging Forensic Detection(2019-09-05) SANTOS, G. L.; OLIVEIRA, G. R.; PRADO, F. F.; SERIKAKU, R.; SANTOS, R. M.; LOPES, G.; RODRIGUES, P. S.© 2019 IEEE.The increasing number of digital media users, as well as the development of multimedia platforms, such as smartphones and tablets, has also increased the number of users who professionally and fraudulently manipulate all types of digital media. This paper proposes a methodology based on an image processing pipeline to detect mainly copy-move type frauds, which are intended to hide or enlarge visual information. Our proposal has been tested in a Copy-Move Forgery database and the results were equal or better in performance than the state-of-art methods.
- A Bio-Inspired Strategy for 3D Surface Reconstruction of Unstructured Scenes Applied to Medical Images(2019-09-05) BOUZON, M.; ALBERTINI, G.; VIANA, G.; MEDEIROS, G.; Paulo Rodrigues© 2019 IEEE.The use of 3D reconstruction, along with immersive technologies, is a technique used in several areas of research and development. Currently, the most common strategy for performing this type of reconstruction is using a stereoscopic camera model. The problem worsens when the challenge involves unstructured scenes, which are scenes that have an ill-defined cognitive architecture. The present work proposes a methodology for 3D reconstruction of unstructured surfaces using monocular cameras. Thus, modern AI techniques, Computer Vision and Computer Graphics techniques have been applied to solve this problem. The experiments performed in this work can be concluded that the proposed method can reconstruct structured scenes with a hit rate between 63% and 68%, depending on the number of thresholds used in the segmentation, thus being superior to the classical method, where the extraction of points is done over the original image without any pre-processing.
Artigo A study of a multi-thresholding segmentation algorithm based on bio-inspired metaheuristic and non-extensive tsallis statistics(2019) RODRIGUES, Paulo; BOUZON, M. F.; HORVATH, M.; VARELA, V. P.; LOPES, GuilhermeAn image segmentation process is one of the most important steps in an image recognition or analysis application pipeline. It is a step that splits each image into disjointed regions of interest. It is also a task that is usually performed by biological processes, such as human visual system. Due to the low processing and ease of implementation, one of the most used techniques is the thresholding method, which consists in finding the best cutting thresholds of a probability distribution histogram. However, the higher the number of thresholds, the greater the computational complexity. And there is no consensus on the number of thresholds and the partitioning position as well. This paper presents a study of the number of thresholds for segmenting an image into their regions of interest. For this purpose, the proposed method uses a bio-inspired algorithm based on meta-heuristics, called firefly with a non-extensive Tsallis statistics kernel. Also, the images are pre-filtered with a low-pass filter based on a q-gaussian function. Using a manually segmented database, the results show that there is an inverse correlation between the Fourier spectrum of an image and the number of thresholds which most approximates the image from the used ground truth. This suggests an automatic method for calculating the required number of thresholds.- A strategy based on non-extensive statistics to improve frame-matching algorithms under large viewpoint changes(2019) LOPES, Guilherme; HORVATH, M.; GIRALDI, G. A.; LOPES, Guilherme