A perceptual grouping and fuzzy logic approach for object recognition from ambiguous images

In this paper, we propose a perceptual grouping approach that uses both geometric relations and fuzzy logic reasoning to develop a computer vision system with the ability to recognize shapes from ambiguous images. The new method takes advantage of fuzzy logic in modeling imprecise systems for reasoning and decision-making, and has shown to be able to solve the shape recognition problem very effectively. Compared to other methods, our approach addresses the ambiguity of image data due to the lack of complete information about the objects under an uncertain environment. In addition, our technique can considerably reduce the computational complexity and data requirement, and hence, is more flexible in real-time applications.