Showing papers by "Jose C. M. Bermudez published in 2022"
TL;DR: In this paper , Coupled tensor LL1-based recovery and blind unmixing of the unknown super-resolution image is used to account for spectral variability in super resolution images.
Abstract: Hyperspectral Super-resolution Accounting for Spectral Variability: Coupled Tensor LL1-Based Recovery and Blind Unmixing of the Unknown Super-resolution Image
10 citations
TL;DR: The results show that a simple neural network structure can lead to a performance equivalent to that of much more complex structures, which are routinely used in the literature.
Abstract: Epithelial dysplasia (ED) is one of the most important factors in detecting the progression of an oral tissue alteration towards carcinoma. Its early detection is instrumental in avoiding the progression of the tumor. A major difficulty for detecting ED is the recognized variability of pathologist assessments. This study proposes a new method that leverages the pathologist expertise to design a simple and efficient classification system to support the detection of dysplastic epithelia. We employ a multilayer artificial neural network (MLP-ANN) and defining the regions of the epithelium to be assessed based on the knowledge of the pathologist. The performance of the proposed solution was statistically evaluated. The implemented MLP-ANN presented an average accuracy of $$87\%$$ , with a variability much inferior to that obtained from three trained evaluators. Moreover, the proposed solution led to results which are very close to those obtained using a convolutional neural network (CNN) implemented by transfer learning, with 100 times less computational complexity. In conclusion, our results show that a simple neural network structure combined with the pathologist expertise can lead to a performance equivalent to that of much more complex structures, which are routinely used in the literature.
TL;DR: The need and the difficulties for the development of data processing methods that are objective, impartial, transparent, explainable, simple to implement and with low computational cost are discussed, aiming to the implementation of risk-based regulation in the world.
Abstract: Access to data and data processing, including the use of machine learning techniques, has become significantly easier and cheaper in recent years. Nevertheless, solutions that can be widely adopted by regulators for market monitoring and inspection targeting in a data-driven way have not been frequently discussed by the scientific community. This article discusses the need and the difficulties for the development of such solutions, presents an effective method to address regulation planning, and illustrates its use to account for the most important and common subject for the majority of regulators: the consumer. This article hopes to contribute to increase the awareness of the regulatory community to the need for data processing methods that are objective, impartial, transparent, explainable, simple to implement and with low computational cost, aiming to the implementation of risk-based regulation in the world.
TL;DR: In this article , the authors studied the behavior of the diffusion LMS (DLMS) algorithm without the assumption of zero delay and found that delays in probing the unknown system yield a bias in the algorithm without increasing its convergence time.
Abstract: Available analyses of the diffusion LMS (DLMS) algorithm assume that the nodes probe the unknown system with zero delay. This assumption is unrealistic, since the unknown system is usually distant from the nodes. The present paper studies the behavior of the algorithm without this assumption. The analysis is done for a network having a central combiner. This structure reduces the dimensionality of the resulting stochastic models while preserving important diffusion properties. Communication delays between the nodes and the central combiner are also considered in the analysis. The analysis is done for system identification for cyclostationary white Gaussian nodal inputs. Mean and mean-square behaviors of the algorithm are analyzed. It is found that delays in probing the unknown system yield a bias in the algorithm without increasing its convergence time. The communication delays between the nodes and the central combiner increase the convergence time without affecting the steady-state behavior. The stability of the algorithm is not affected by either type of delay. The analysis exactly matches the simulations.
01 Nov 2022