Canonical frames for planar object recognition

Conference item

We present a canonical frame construction for determining projectively invariant indexing functions for non-algebraic smooth plane curves. These invariants are semi-local rather than global, which promotes tolerance to occlusion.

Two applications are demonstrated. Firstly, we report preliminary work on building a model based recognition system for planar objects. We demonstrate that the invariant measures, derived from the canonical frame, provide sufficient discrimination between objects to be useful for recognition. Recognition is of partially occluded objects in cluttered scenes. Secondly, jigsaw puzzles are assembled and rendered from a single strongly perspective view of the separate pieces. Both applications require no camera calibration or pose information, and models are generated and verified directly from images.

Authors: Rothwell, CA; Zisserman, A; Forsyth, DA; Mundy, JL

• Publication status: Published
• Peer review status: Peer reviewed
• Publisher: Springer
• Host title: Computer Vision — ECCV ’92 Second European Conference on Computer Vision Santa Margherita Ligure, Italy, May 19–22, 1992 Proceedings
• Pages: 757-772
• Series: Lecture Notes in Computer Science
• Series number: 588
• Publication date: 1992-04-29
• Event title: 2nd European Conference on Computer Vision (ECCV 1992)
• Event location: Santa Margherita Ligure, Italy
• Event start date: 1992-05-19
• Event end date: 1992-05-22
• DOI: 10.1007/3-540-55426-2_86
• ISSN: 1611-3349
• EISBN: 9783540470694
• ISBN-10: 3540554262
• ISBN-13: 9783540554264
• Language: English