| Active models have been widely used in image segmentation. Typical roadblocks to consistent performance include limited capture range, noise sensitivity, poor convergence to concavities, expensive computing cost, and tedious initialization. This dissertation proposes a new external force for active models, called vector field convolution (VFC), and a novel automated initialization method, termed Poisson inverse gradient (PIG) initialization, to address these problems.; VFC is computed by convolving an edge map generated from the image with a user-defined vector field kernel. We propose two structures for the magnitude function of the vector field kernel, and we provide an analytical method to estimate the parameter of the magnitude function. A modified version of VFC is introduced to alleviate the possible leakage problem caused by choosing inappropriate parameters. We also demonstrate that the classical external force and the gradient vector flow (GVF) external force are special cases of VFC in certain scenarios. Examples and comparisons with GVF are presented in this dissertation to show the advantages of this innovation, including large capture range, convergence to concavities, superior robustness to noise, reduced computational cost, and the flexibility of tailoring the external force.; A crucial stage that affects the ultimate active model performance is initialization. This dissertation proposes a novel automatic initialization approach for active models in both 2D and 3D. The PIG initialization method exploits a novel technique that essentially estimates the external energy from the external force and determines the most likely coarse segmentation. Examples and comparisons with two state-of-the-art automatic initialization methods are presented in this dissertation to illustrate the advantages of this innovation, including the ability to choose the number of active models deployed, rapid convergence, accommodation of broken edges, superior noise robustness and segmentation accuracy.; Index Terms---Image segmentation, active models, deformable models, snakes, active contours, deformable contours, active surfaces, deformable surface, external force, initialization. |
