HK segmentation of 3D micro-structures reconstructed from focus

Official URL: http://dx.doi.org/10.1007/978-3-540-87442-3_145

This paper presents an evaluation of HK segmentation on 3D micro-structures reconstructed from focus. Due to the necessity of 3D shape and surface structure information for a precise micromanipulation, shape of the object is recovered using shape from focus (SFF). In the SFF procedure, using an image sequence, composed of images captured at different focusing levels, focused image of the micro-structure and its shape are acquired. Then the resulting shape, also called range image, is used in HK segmentation method to extract surface curvature information. Experimental results show that this technique works for both synthetic and real data.