next up previous
Next: Bibliography Up: Moment Representation of Blobs Previous: 3. Practice

   
4. Discussion

With a scene under careful control, with excellent fixed lighting, and with limited target classes, the method of moment invariants can form the basis for a translation, rotation and scale independent machine vision system. The blob silhouette and edge images used in the preceding work may be impractical for all but the most constrained problems, as even some touch up prior to processing using an image pixel editing program was performed by the author. However, there exist very few machine vision algorithms which would execute faster than simple moment calculations, especially while maintaining independence from four degrees of freedom!


next up previous
Next: Bibliography Up: Moment Representation of Blobs Previous: 3. Practice
Mike Andrews
1999-04-09