Identification of 3D objects using multicolored grid coding

✍ Scribed by Ayako Takenouchi; Masato Nakajima

• Publisher: John Wiley and Sons
• Year: 1998
• Tongue: English
• Weight: 195 KB
• Volume: 29
• Category: Article
• ISSN: 0882-1666
• DOI: 10.1002/(sici)1520-684x(199804)29:4<60::aid-scj6>3.0.co;2-m

✦ Synopsis

A technique has been developed for encoding the image of a 3D object from two viewpoints. Using doublecolored grid coding and a neural network, a wider variety of 3D objects can be identified. The image is encoded as a modulated grid pattern by projecting a grid pattern over the object. Using two simultaneous views enables more information to be obtained about the object. Features of the object that are invariant to its position and orientation are extracted from the spectrum of the coded image and are input to a neural network, which performs object recognition. Experiments on simple objects gave an identification rate of about 95%.