Codon Constraints on Closed 2D Shapes,
MASSACHUSETTS INST OF TECH CAMBRIDGE ARTIFICIAL INTELLIGENCE LAB
Pagination or Media Count:
Codons are simple primitives for describing plane curves. They thus are primarily image-based descriptors. Yet they have the power to capture important information about the 3D world, such as making part boundaries explicit. The codon description is highly redundant useful for error-correction. This redundancy can be viewed as a constraint on the number of possible codon strings. For smooth closed strings that represent the bounding contour silhouette of many smooth 3D objects, the constraints are so strong that sequences containing 6 elements yield only 33 generic shapes as compared with a possible number of 15,625 combinations. An important task for object recognition is the description of the shape of a bounding contour such as a sihouette that outlines as object. Although recognition need require only partial segments of such contours, the internal canoncial description, against which the image contour is compared, is very likely a closed ring. Our concept of most objects should lead us to expect such a closed contour. The description of closed, 2D contours thus is an important ingredient of a system for object recognition. First the author present such a scheme, described in more detail elsewhere and then show how the scheme leads to a hierarchical taxonomy of closed, 2D shapes. Additional keywords Image understanding Shape representation Applied mathematics Artificial intelligence.
- Theoretical Mathematics