Document Type
Thesis
Date of Award
9-30-1990
Degree Name
Master of Science in Mechanical Engineering - (M.S.)
Department
Mechanical and Industrial Engineering
First Advisor
Rajesh N. Dave
Second Advisor
Ian Sanford Fischer
Third Advisor
Nouri Levy
Abstract
The problem of shape detection has been treated as one of clustering. A fuzzy-shell clustering(FSC) has been formulated as non linear least squares with fuzzy memberships as weights. An improved FSC algorithm has been proposed based on this formulation to partition the input feature vectors into hyper-spheres. To illustrate the usefulness of improved FSC we have provided numerous examples in 2-D space. The idea has been extended to fuzzy ellipsoidal-shell clustering (rESC). A FESC algorithm has been proposed to partition the input feature vectors into ellipsoidal patterns. To substantiate our claims regarding usefulness of this algorithm we have provided many examples in 2-D space.
Recommended Citation
Patel, Kalpesh J., "Characteristic shape recognition with fuzzy dustering" (1990). Theses. 2889.
https://digitalcommons.njit.edu/theses/2889