TY  - CONF
AU  - Gioacchino Vino
AU  - Angel Sappa
A2  - ICIAR
PY  - 2013//
TI  - Revisiting Harris Corner Detector Algorithm: a Gradual Thresholding Approach
T2  - LNCS
BT  - 10th International Conference on Image Analysis and Recognition
SP  - 354
EP  - 363
VL  - 7950
PB  - Springer Berlin Heidelberg
N2  - This paper presents an adaptive thresholding approach intended to increase the number of detected corners, while reducing the amount of those ones corresponding to noisy data. The proposed approach works by using the classical Harris corner detector algorithm and overcome the difficulty in finding a general threshold that work well for all the images in a given data set by proposing a novel adaptive thresholding scheme. Initially, two thresholds are used to discern between strong corners and flat regions. Then, a region based criteria is used to discriminate between weak corners and noisy points in the midway interval. Experimental results show that the proposed approach has a better capability to reject false corners and, at the same time, to detect weak ones. Comparisons with the state of the art are provided showing the validity of the proposed approach.
SN  - 0302-9743
SN  - 978-3-642-39093-7
UR  - http://dx.doi.org/10.1007/978-3-642-39094-4_40
N1  - ADAS; 600.055
ID  - Gioacchino Vino2013
ER  -