Object detection with feature stability over scale space

This paper proposes a novel segmentation method based on the scale space techniques endowed with a feature stability approach. The novelty of the paper is the lifetime of the space-scale blobs measured not only by their presence and disappearance but by the stability of the features characterizing the objects of interest as well. Our numerical experiments show that the algorithm outperforms the conventional space scale algorithm applied to variable size and variable shape objects. The proposed algorithm can be used as a preprocessing step in object or pattern recognition applications to produce seeds for more accurate image segmentation methods such as the snakes or the level set techniques.

► We propose a segmentation method based on scale space endowed with feature stability.
► The stability of blob’s features is taken into account for blob lifetime measurement.
► The algorithm outperforms conventional space scale algorithm.
► Variable size and variable shape objects are efficiently detected by the algorithm.
► It can also be used as a preprocessing step in pattern recognition applications.