Applying Ternary Net Weights to the R-Mask Neural Network to Identify Bronchopulmonary Lung Segments

Autor:	N. J. Francis, S. A. Aljasar, Yubin Xu, S. V. Axyonov, N. S. Francis, Muhammad Saqib
Rok vydání:	2020
Předmět:	History Artificial neural network Computer science Activation function ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION Net (mathematics) Hyperbolic tangent function Computer Science Applications Education Lung segmentation Tomography Graphics Ternary operation Algorithm
Zdroj:	Journal of Physics: Conference Series. 1611:012061
ISSN:	1742-6596 1742-6588
DOI:	10.1088/1742-6596/1611/1/012061
Popis:	The purpose of this research is to develop an algorithm for detecting bronchopulmonary segments in lung Computer Tomography (CT) images, while reducing computational costs. The algorithm is implemented without the use of a graphics processor (GPU). The main algorithm of the proposed system introduces ternary weights into Mask R-CNN. The ternary hyperbolic tangent function replaces Mask R-CNN’s activation function to reduce overhead costs. This is a convenient and inexpensive system, designed to help radiologists to detect bronchopulmonary lung segmentation with high accuracy.
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=doi_________::fc46f22aae4934b4e8d9a10e3128825e https://doi.org/10.1088/1742-6596/1611/1/012061 Zobrazit plný text záznamu