Multi-layer Lacunarity for Texture Recognition

摘要

Lacunarity could be applied for the analysis of different types of textures. Application of binary lacunarity for the analysis of grayscale images is proposed in this paper. Input image is thresholded using predefined set of values and every binary image is processed using lacunarity for the selected scale. Obtained vector could be used for the analysis and segmentation purposes. Achieved probability of identification is 30% for ideal detection and more then 60% for 10% acceptance margin.