Bilateral Kernel-Based Region Detector
Woon Cho, Sung Yeol Kim, Andreas F. Koschan, Mongi A. Abidi · 2013
In this paper, we present a new method for a locally adaptive region detector called Bilateral kernel-based Region Detector (BIRD). This work is to detect stable regions from images by consecutively computing a multiscale decomposition based on the bilateral kernel. The BIRD regards a region as covariant if it exhibits predictability in its photometric distance over spatial distance. Distinctiveness and robustness across scales are achieved by selecting the extremely stable regions through sequential scales. Our method is simple and easy to implement. Experimental results show that our method outperforms competing affine region detection methods in efficiency on region detection. watershed segmentation algorithm to the image and extracts homogeneous intensity regions over a large range of thresholds. Tuytelaars and Gool [9] proposed intensity-based regions (IBR) and edge-based regions (EBR). Figure 1 illustrates the results for several detectors to demonstrate the interpretability and complementarity on an image of a Siemens star with regular beams as mentioned in [3]. MSER stably extracts a region for each beam in perfect symmetry, and no non-interpretable features at all. Contrarily results for Haraff, Hesaff, IBR, and EBR are neither symmetric nor easy to interpret. 1.