Hands: a pattern theoretic study of biological shapes
Hands: a pattern theoretic study of biological shapes
Object Matching Using Deformable Templates
IEEE Transactions on Pattern Analysis and Machine Intelligence
On shape detection in noisy images with particular reference to ultrasonography
Statistics and Computing
Parts-based segmentation with overlapping part models using Markov chain Monte Carlo
Image and Vision Computing
| Hi-index | 0.00 |
This article addresses the problem of identification of partly destroyed human melanoma cancer cells in confocal microscopy imaging. Complete cancer cells are nearly circular and most of them have a nearly homogeneous boundary and interior region. A deformable template (Grenander, 1993) is well suited for these complete cells and models a cell as a natural deformed template or prototype. We will in this article focus on the remaining cells which have lost parts of the boundary region most probably due to a ’capping‘ phenomenon. We can interpret these cells as being partly destroyed, where in our statistical model the lost part of the boundary region is generated by a destructive deformation field acting and living on the cell or template. By doing simultaneous inference for both the natural and destructive deformation field, we are able to obtain reliable estimates of the outline in addition to where on the boundary the cell is destroyed. We apply our model to identifying partly destroyed human melanoma cancer cells with good results.