Abstract | ||
---|---|---|
We present a statistical method that leads to accurate volume measurements of lung tumors from computerized tomographic (CT) data. The method is based on the assumption that a range of pixel intensities in CT data defines the edge of a tumor, and from our statistical model, we assign a probability that a given pixel intensity is included in the tumor measurement. Using the magnitude of the gradient of the pixel intensities over the density range observed for lung tumors and lung tissue, we have found consistent metrics that help define these weights, so that the measurement does not require user-controlled parameters and can be performed automatically. This could ultimately lead to direct comparisons of measurements from different medical laboratories. |
Year | Venue | Keywords |
---|---|---|
2009 | IPCV | robust,image processing,segmentation,biweight,statistical model |
Field | DocType | Citations |
Computer vision,Magnitude (mathematics),Pattern recognition,Segmentation,Computer science,Image processing,Pixel,Statistical model,Artificial intelligence | Conference | 4 |
PageRank | References | Authors |
0.74 | 8 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Adele P. Peskin | 1 | 33 | 6.79 |
Karen Kafadar | 2 | 46 | 6.32 |
A. M. Santos | 3 | 4 | 0.74 |
Gillian Haemer | 4 | 4 | 0.74 |