원문정보
초록
영어
A novel operator is presented which is suitable for object detection in images. The operator creates a so called area map of an image. Each element in this area map is calculated by evaluating an indicator function which computes the size of image areas according to predened and problem specic properties. In general, results are approximate, so further renement and verication is necessary for a reliable outcome of object detection. In this work the area map approach is categorized in relation to classical image processing algorithms and the dierence to other approaches is pointed out. Two specic area maps, the DecAMap (decrementing area map) and the IncAMap (incrementing area map) are used to illustrate the principle of operation on an example of indentation detection in Vickers hardness testing images.
목차
1: Introduction
2: AreaMap operator
2.1: De nition and classi cation
2.2: Delimitation from other object detection algorithms
3: Customizing AreaMap for Vickers indentation images
3.1: Indentation image properties
3.2: AreaMap performance optimization
3.3: Indentation center determination
4: Experiments
5: Conclusion
Acknowledgements
References