Laplante P A, Sinha D
Technol. Educ. Center, New Jersey Inst. of Technol., Newark, NJ.
IEEE Trans Syst Man Cybern B Cybern. 1996;26(1):21-8. doi: 10.1109/3477.484435.
In all man-machine systems with image processing functions, an important unsolved problem arises in the treatment of uncertain and incomplete image information. Several frameworks have been suggested for handling uncertain image information including; expert systems, fuzzification, likelihood estimation, and neural networks. In this paper we review those methods. We also present a new method for handling uncertainties by unifying the representations of gray-values and uncertainty into one framework in a way that parallels fuzzy logic. This new framework is based on the application of the extended fuzzy pointed set and an associated algebra to handle uncertain information. We further show how this framework can be used in image processing and artificial intelligence.
在所有具有图像处理功能的人机系统中，在处理不确定和不完整的图像信息时都会出现一个重要的未解决问题。已经提出了几种处理不确定图像信息的框架，包括专家系统、模糊化、似然估计和神经网络。在本文中，我们回顾了这些方法。我们还提出了一种新的方法来处理不确定性，即将灰度值和不确定性的表示统一到一个框架中，其方式类似于模糊逻辑。这个新框架基于扩展模糊点集和相关代数的应用来处理不确定信息。我们进一步展示了这个框架如何用于图像处理和人工智能。
