På svenska

Presentation Information     2009-10-05 (15:15)   •  The seminar room at Vi2

Speaker Joakim Lindblad
Title Pixel coverage segmentation for improved feature estimation
Abstract To make precise measurements of object features based on digital images is at the core of almost any Image Analysis application. In a number of publications it has been shown that significant improvements in precision of object feature estimates can be achieved by properly preserving and utilizing intensity information available in images. We have begun the development of a framework for high precision feature estimates based on pixel coverage representations. To enable application of this framework to real images, appropriate segmentation methods are required. In this talk four approaches for pixel coverage segmentation will be briefly presented. One of the methods (recently presented at ICIAP) will be described in more detail, and its use will be exemplified on histological images where it enables precise measurements of bone implant integration.

In addition you may also be shown some photos from the ICIAP and ISPA conferences.