A fast algorithm for level set-like active contours
Pattern Recognition Letters
An Experimental Comparison of Min-Cut/Max-Flow Algorithms for Energy Minimization in Vision
IEEE Transactions on Pattern Analysis and Machine Intelligence
ZIB structure prediction pipeline: composing a complex biological workflow through web services
Euro-Par'06 Proceedings of the 12th international conference on Parallel Processing
Automatic Segmentation of High-Throughput RNAi Fluorescent Cellular Images
IEEE Transactions on Information Technology in Biomedicine
| Hi-index | 0.00 |
The segmentation of complex touching and overlapping cells in fluorescent micrographs poses a challenge for automated image analysis systems. In order to improve performance for complex image data, multi-channel approaches exist that additionally incorporate information from the cell nuclei. The most frequently method used for fluorescent micrograph segmentation is the seeded watershed transform. But methods based on level sets and graph cuts can also be used as alternatives to the watershed transform based splitting of cells. In this work we investigate if segmentation results obtained by one of the named methods are superior in terms of segmentation performance. Therefore, a hybrid watershed transform based method, a very efficient fast marching cell splitting method and a cell splitting method based on graph cuts are described and investigated. For performance comparison the parameters of each method are automatically optimized toward a manual ground truth including cross validation techniques. Our evaluations show that for the presented dataset of bone marrow-derived macrophages the hybrid watershed transform based method can compete with both, the fast marching level set and the graph cut based method.