Grant: Discovery Grants Program - Individual

The general objective of this proposed research program is to discover the theorems, algorithms, and solutions for a number of issues related to the moment descriptors in the field of image analysis. Since the moments of an image can uniquely provide global characteristics and geometrical features of the image with some invariant properties, the moment methods have been utilized in many applications in image analysis, pattern recognition, object classification, and many other scientific fields in recent years. However, the theoretical supports on some fundamental issues have not attracted sufficient attention, and the lack of such supports has limited the role that moment descriptors are able to play. These issues include: (1) ) The need to discover new types of image moments to enrich the moment methods; (2) ) The prospect of applying moment methods to images stained by various types of noise; (3) ) The relationship between global moments and local moments of the same image. This proposed research intends to undertake the abovementioned three outstanding issues. To resolve these subjects, the new theorems, algorithms, and techniques are needed to be developed. The answers to these inquiries will undoubtedly have strong impacts on the moment-based applications in image analysis and other scientific fields.