The Design and Use of Steerable Filters
IEEE Transactions on Pattern Analysis and Machine Intelligence
Inferring global perceptual contours from local features
International Journal of Computer Vision - Special issue on computer vision research at the University of Southern California
A Parametric Texture Model Based on Joint Statistics of Complex Wavelet Coefficients
International Journal of Computer Vision - Special issue on statistical and computational theories of vision: modeling, learning, sampling and computing, Part I
Coherence-Enhancing Diffusion Filtering
International Journal of Computer Vision
Deformable Kernels for Early Vision
IEEE Transactions on Pattern Analysis and Machine Intelligence
Constrained Flows of Matrix-Valued Functions: Application to Diffusion Tensor Regularization
ECCV '02 Proceedings of the 7th European Conference on Computer Vision-Part I
Tensor Voting: A Perceptual Organization Approach to Computer Vision and Machine Learning (Synthesis Lectures on Image, Video, and Multimedia Processing)
International Journal of Computer Vision
Design of steerable filters for feature detection using canny-like criteria
IEEE Transactions on Pattern Analysis and Machine Intelligence
Steerable wedge filters for local orientation analysis
IEEE Transactions on Image Processing
IEEE Transactions on Image Processing
| Hi-index | 0.00 |
Steerable filters are a common tool for feature detection in early vision. Typically, a steerable filter is used as a matched filter by rotating a template to achieve the highest correlation value. We propose to use the steerable filter bank in a different way: it is interpreted as a model of the image formation process. The filter maps a hidden 'orientation' image onto an observed intensity image. The goal is to estimate the hidden image from the given observation. As the problem is highly under-determined, prior knowledge has to be included. A simple and effective regularizer which can be used for edge, line and surface detection will be used. Further, an efficient implementation in terms of Circular Harmonics in the conjunction with the iterated use of local neighborhood operators is presented. It is also shown that a simultaneous modeling of different lowlevel features can improve the detection performance. Experiments show that our approach outperforms other existing methods for low-level feature detection.