| 摘要 |
<p>The systems described herein employ a pre-processing strategy for image databases to obtain derivative information from images, in particular from within regions of interest (if any) in each image. A similar pre-processing strategy may be applied to a query image so that rapid and highly accurate comparisons may be performed between the query image and images from the image database, based upon the derivative information. The pre-processing strategy may include, for example, normalization of image data, sampling of normalized image data, segmentation of regions of interest, and extraction of feature vectors from the normalized, sampled, segmented image data. The overall strategy may provide for efficacious comparison of new and archived images. Further disclosed herein are techniques employing these strategies in modeling normalized image data for various applications, in organizing and searching databases, in providing a user interface for applying the systems in a medical workplace, and for deploying applications based upon these strategies locally and over computer networks.</p> |
