The Application of Local Structure Classification in Content-Based 3D Medical Image Retrieval

Abstract

An approach is proposed to classify local structures for 3D medical images into line-like, blob-like and sheet-like patterns in different sizes. This approach is based on the second-order derivatives in multiscale space, using normalized response functions. The application of this approach in content-based 3D medical image retrieval is discussed. Experiments on CT images are performed and the method is proved to be efficient for filtering specific local structures in selected range of sizes.

DOI: 10.1109/ICIG.2013.132

4 Figures and Tables

Cite this paper

@article{Dong2013TheAO, title={The Application of Local Structure Classification in Content-Based 3D Medical Image Retrieval}, author={Shuzhao Dong and Kaikun Dong and Lu Yin}, journal={2013 Seventh International Conference on Image and Graphics}, year={2013}, pages={639-642} }