Log-Polar transform in 3D environment
Takayuki Nakata, Yue Quan Bao · 2008
Log-Polar transform is mentioned as one of the models of the central-fovea visual sensor in a creature sight. Up to now, in the field of image recognition, Log-Polar transform is known widely and is especially used well in recognition of face images. However, since it is the recognition method only to a scaling and a rotation, it mainly remains in application in a 2D space. In this paper, a new Log-Polar transform algorithm which can detect even affine transform parameters is proposed, and this algorithm is expand so that perspective transform invariant can be detected. By using this transform, the pattern recognition within 3D space is possible.