IVADepending on the application, local feature detectors should comply with properties that are often contradictory, e.g. distinctiveness vs. robustness. Providing a good balance is a standing problem in the field. In this direction, we propose a novel approach for \emph{local feature detection} starting from sampled edges and based on shape stability measures across the \emph{weighted a-filtration}, a computational geometry construction that captures the shape of a non-uniform set of points. Detected features are blob-like and include non-extremal regions as well as regions determined by cavities of boundary shape. The detector provides distinctive regions, while achieving high robustness in terms of \emph{repeatability} and \emph{matching score}, as well as competitive performance in a large scale image retrieval application.
 |
 |
 |
 |
| (a) | (b) | (c) | (d) |
Figure 1. Feature detection based on a-shapes. (a) Input image. (b) Edge samples with assigned weights as squared radii (blue) and triangulation. Triangle colors denote distinct connected components of the a-complex. Detected features by: (c) WαSH detector and (d) MSER.
WαSH detection results on images from the bikes and Oxford Buildings datasets.
 |
 |
 |
 |
Repeatability, matching score and number of correct matches, on the bikes, boat, lauven and wall datasets.
 |
 |
 |
 |
 |
 |
 |
 |
 |
Win32 Binary file
Linux 32bit binary file
Linux 64bit binary file