Restoring images with subpixel precision using convex restrictions Amauri Antunes Filho, Murillo Rodrigo Petrucelli Homem ACM International Conference Proceeding Series, 2017 The objective of Super-Resolution is to obtain an image with higher resolution by using the information coming from one or more low resolution images. There are many different applications where super resolution is employed. Medical and forensic are some examples. This work is aimed at studying and developing algorithms based on Teklap and Sezan algorithms that use the theory of projection over convex sets to obtain a higher resolution image based on a set of images with subpixel information. We also propose a convex restriction based on Richardson-Lucy algorithm modified to weight Canny filter together with total variation regularization aiming at restoring lost frequencies in order to have a hybrid approach in implementing a spatial and spectral super-resolution simultaneously. The results present improvement in the images obtained compared to low resolution images, minimizing noise and blur and improving definition of edges. It can be concluded that this approach has a great potential to be used in different applications such as medical and forensic images.
An innovative augmented reality educational framework with gamification to assist the learning process of children with intellectual disabilities Rogerio Colpani, Murillo Rodrigo Petrucelli Homem Iisa 2015 6th International Conference on Information Intelligence Systems and Applications, 2016 Currently, several studies are making use of multimedia systems, and Virtual Reality (VR) technology has been applied to people with special needs. However, its main limitations are the need for qualified human resources and the high costs. On the other hand, Augmented Reality (AR) technology has been increasing and it has become more and more popular because of its specificities. However, most studies involving these technologies are focused on the treatment of people with motor disabilities. Thus, this paper presents a proposal of an AR framework with gamification to assist the learning process of children with intellectual disabilities in general. Finally, the study will present some ways on how teachers might work some concepts and cognitive skills on children with intellectual disabilities with the aid of the framework.
Preliminary study on systematic literature review of augmented and virtual reality applied to motor and cognitive disabilities Rogerio Colpani, Murillo Rodrigo Petrucelli Homem 2015 10th Iberian Conference on Information Systems and Technologies Cisti 2015, 2015 In recent years, studies have been making use of Virtual and Augmented Reality technologies, which has shown promising results. However, there are few studies conducted on the disabilities area. From this approach, this paper presents the results of a preliminary study on systematic literature review of augmented and virtual reality applied to motor and cognitive disabilities.
Super-resolution image reconstruction using the generalized isotropic multi-level logistic model Ana L. D. Martins, Murillo R. P. Homem, Nelson D. A. Mascarenhas Proceedings of the ACM Symposium on Applied Computing, 2009 High spatial resolution images are usually required in a great number of applications such as video surveillance, for instance. Super-Resolution reconstruction methods use image processing techniques to estimate a high-resolution image based on a set of low-resolution observations of the same scene. Therefore, these methods are able to overcome cost and hardware limitations inherent to acquisition devices. This paper discusses a Maximum a Posteriori Probability approach, characterizing the high-resolution estimation with the Isotropic Multi-Level Logistic Model that incorporates pixel similarity in a meaningful way to the super-resolution context. Following, the high-resolution estimation is derived by maximizing the local conditional probabilities sequentially with the Iterated Conditional Modes algorithm. The proposed method was evaluated in a simulated framework using the Normalized Mean Square Error criterion, and in a real situation using video frames. The results indicate the effectiveness of our approach both by numerical and visual evaluation.
MAP-MRF super-resolution image reconstruction using maximum pseudo-likelihood parameter estimation Ana L. D. Martins, Alexandre L. M. Levada, Murillo R. P. Homem, Nelson D. A. Mascarenhas Proceedings International Conference on Image Processing Icip, 2009 In this paper, we address the parameter estimation of a Super-Resolution Image Reconstruction approach following a Maximum a Posteriori Probability (MAP) algorithm. The Generalized Isotropic Multi-Level Logistic (GIMLL) Markov Random Field (MRF) model is considered for the high-resolution image characterization. In most applications, MRF model parameters are still chosen by a trial-and-error procedure through simple manual adjustments. In order to overcome this problem we propose a novel approach based on interval parameter estimation using both the Maximum Pseudo-Likelihood (MPL) technique and an approximation of the asymptotic variance of this estimator. To evaluate the capability of the proposed estimator we used a Markov Chain Monte Carlo algorithm to generate GIMLL model outcomes. The differences between the real parameters and the proposed MPL estimators are not significant. Moreover, the Normalized Mean Square Error (NMSE) of the high-resolution estimations indicate the effectiveness of our approach and the importance of an accurate estimation procedure.
Super-resolution image reconstruction using the ICM algorithm A. L. D. Martins, M. R. P. Homem, N. D. A. Mascarenhas Proceedings International Conference on Image Processing Icip, 2007 Super-resolution image reconstruction is a powerful methodology for resolution enhancement from a set of blurred and noisy low-resolution images. Following a Bayesian framework, we propose a procedure for super-resolution image reconstruction based on Markov random fields (MRF), where a Potts-Strauss model is assumed for the a priori probability density function of the actual image. The first step is given by aligning all the low-resolution observations over a high-resolution grid and then improving the resolution through the iterated conditional modes (ICM) algorithm. The method was analyzed considering a number of simulated low-resolution and globally translated observations and the results demonstrate the effectiveness of the algorithm in reconstructing the desirable high-resolution image.
Skeletonization of two-dimensional shapes via fast numerical calculation of vector fields Brazilian Symposium of Computer Graphic and Image Processing, 2004
Linear filters for deconvolution microscopy Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation, 2004
Method for nondestructive testing using multiple energy CT and statistical pattern classification Proceedings of SPIE the International Society for Optical Engineering, 1999
