Ear segmentation using histogram based K-means clustering and Hough transformation under CVL dataset
Heng Liu, Dekai Liu · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2009
Under CVL dataset, we provide an image segmentation approach based on adaptive histogram based K-means clustering and fast Hough transformation. This work firstly analyzes the characteristics of ear images in CVL face dataset. According to the analysis, we then use adaptive histogram based K-means clustering method to threshold ear images and then roughly segment the ear parts. After ear contour extraction, with boundary determination through vertical project, Hough transformation is utilized to locate the ear contour accurately. The experimental results and comparisons with other segmentation methods show our approach is effective.