Develop control point suppression techniques for very large networks
This is a new application. No impact on existing ISIS applications.
Using the findfeatures application can produce control networks that are very large. They may contain a very large number of control points/image-based control network with many image measures/point (depth).
These networks become difficult to processing due to their size and volume. An application (cnetsuppress) will be developed to reduce/suppress the number of control points in networks while computing the most efficient distribution of the best registered control points with an emphasis on maintaining point (measure) depth.
We intend to use the algorithm described in the paper titled EFFICIENTLY SELECTING SPATIALLY DISTRIBUTED KEYPOINTS FOR VISUAL TRACKING (http://www.cs.ucsb.edu/~holl/pubs/Gauglitz-2011-ICIP.pdf).