Design of steerable filters for feature detection using canny-like criteria

Abstract

Abstract-We propose a general approach for the design of 2D feature detectors from a class of steerable functions based on the optimization of a Canny-like criterion. In contrast with previous computational designs, our approach is truly 2D and provides filters that have closed-form expressions. It also yields operators that have a better orientation selectivity than the classical gradient or Hessianbased detectors. We illustrate the method with the design of operators for edge and ridge detection. We present some experimental results that demonstrate the performance improvement of these new feature detectors. We propose computationally efficient local optimization algorithms for the estimation of feature orientation. We also introduce the notion of shape-adaptable feature detection and use it for the detection of image corners.

Vimentin Intermediate Filaments Template Microtubule Networks to Enhance Persistence in Cell Polarity and Directed Migration

“…Briefly, the raw images are processed with a steerable filter (Jacob and Unser, 2004) to enhance the signal of curvilinear features ( Figure 1D-E), followed by a non-maximum suppression step that skeletonizes the filtered images into line segments ( Figure 1F). Finally, the line segments are aggregated into a complete representation of a filament network using a graph-based matching approach that closes breaks between filaments caused by unevenness of fluorescence signal, low SNR and filament crossovers ( Figure 1G-J, Movies S2).…”

Supervised Feature Learning for Curvilinear Structure Segmentation

“…To generate them, most approaches compute a local linearity measure and threshold the resulting scores. This linearity measure can be postulated a priori [4,5], optimized to find specific patterns [6,7], or learned [8][9][10] from training data.…”

Parameter Estimation for Ridge Detection in Images with Thin Structures

“…The steerable filters [11] are 2D feature detectors, i.e., a class of steerable functions based on the optimization of a Canny-like model. These filters have closedform expressions and lead to operators that have a better orientation selectivity than classical Hessian-based detectors.…”

“…Four techniques are explored: a modified Canny detector, a morphological approach, the Frangi filter (see [7,6]) and a steerable filter [11]. ROC curves are used to analyze these techniques and to find their optimal parameters.…”