Probabilistic reasoning in intelligent systems: networks of plausible inference
Probabilistic reasoning in intelligent systems: networks of plausible inference
Fast Approximate Energy Minimization via Graph Cuts
IEEE Transactions on Pattern Analysis and Machine Intelligence
IEEE Transactions on Pattern Analysis and Machine Intelligence
Comparison of Graph Cuts with Belief Propagation for Stereo, using Identical MRF Parameters
ICCV '03 Proceedings of the Ninth IEEE International Conference on Computer Vision - Volume 2
SIOX: Simple Interactive Object Extraction in Still Images
ISM '05 Proceedings of the Seventh IEEE International Symposium on Multimedia
A comparative evaluation of interactive segmentation algorithms
Pattern Recognition
Nugget-cut: a segmentation scheme for spherically- and elliptically-shaped 3D objects
Proceedings of the 32nd DAGM conference on Pattern recognition
Automatic detection and segmentation of axillary lymph nodes
MICCAI'10 Proceedings of the 13th international conference on Medical image computing and computer-assisted intervention: Part I
IEEE Transactions on Image Processing
| Hi-index | 0.00 |
Automatic segmentation methods for tumors are typically only suitable for a specific type of tumor in a specific imaging modality and sometimes lack in accuracy whereas manual tumor segmentation achieves the desired results but is very time consuming. Interactive segmentation however speeds up the process while still being able to maintain the accuracy of manual segmentation. This paper presents a novel method for fast interactive segmentation of tumors (called FIST) from medical images, which is suitable for all somewhat spherical tumors in any 3d medical imaging modality. The user clicks in the center of the tumor and a belief propagation based iterative adaption process is initiated, thereby considering image gradients as well as local smoothness priors of the surface. During that process, instant visual feedback is given, enabling to intervene in the adaption process by sketching parts of the contour in any cross section. The approach has successfully been applied to the segmentation of liver tumors in CT datasets. Satisfactory results could be achieved in 15.20875 seconds on the average. Further trials on oropharynx tumors, liver tumors and the prostate from MR images as well as lymph nodes and the bladder from CT volumes demonstrate the generality of the presented approach.