Date of Award:
Doctor of Philosophy (PhD)
Connectedness plays an important role in human cognitive and learning activities. Human vision system is very sensitive to the connectedness property of objects and effective in "calculating" the property. The study of connectedness will help us explore the way of human brain extracting the global properties of objects, and enable different avenues to design new artificial intelligent (AI) systems with better performances.
However, connectedness is rarely considered in current AI systems because of the lack of complete theoretic system and efficient computation algorithm. In this work, I focus on building the connectedness theory and algorithms in digital space, and apply them to solve many challenging problems in natural image and low-quality biomedical image segmentation.
The newly proposed Neutro-Connectedness (NC) theory makes it possible to describe the part and the whole relationship and to model the "hidden factors" influencing the decision-making. By applying the dynamic programming strategy, a fast algorithm is proposed to calculate NC for general dataset. It calculates the NC map, and also outputs the NC forest to discover the topological structure of a dataset. The power of NC is demonstrated by applying it to solve two challenging applications: interactive image segmentation (IIS) and breast ultrasound image segmentation (BUSIS).
Xian, Min, "Neutro-Connectedness Theory, Algorithms and Applications" (2017). All Graduate Theses and Dissertations, Spring 1920 to Summer 2023. 6527.
Copyright for this work is retained by the student. If you have any questions regarding the inclusion of this work in the Digital Commons, please email us at .