A Combined Shape-Newton and Topology Optimization Technique in Real-Time Image Segmentation
In this paper, for solving a class of shape optimization problems, a new algorithmic concept combining shape and topological sensitivities is presented. The geometry of interest is represented by means of geometrical implicit functions. While the topology optimization phase is based on a gradient descent scheme, the phase based on shape sensitivity uses Newton-type descent flows. The algorithm is applied to edge detector based image segmentation problems, but it can be extended to solve more general shape optimization problems as well.
Citable link to this pagehttps://hdl.handle.net/1911/102022
MetadataShow full item record
- CAAM Technical Reports