Object segmentation by template matching and its related techniques

International Conference on Nuclear Medicine & Radiation Therapy

July 14-15, 2016 Cologne, Germany

Junyan Wang

University of California Los Angeles, USA

Posters & Accepted Abstracts: J Nucl Med Radiat Ther

Abstract :

Automatic object segmentation is one of the most important tasks in image analysis. While a universal schema for this task is seemingly not shortly attainable, practical solutions under different circumstances have been developed in the past couple of decades. Most of the previous works were based on machine learning. In contrast, our agenda differs from them in that we try to achieve satisfactory segmentation with a limited number of examples. Our idea is to leverage object matching in the segmentation model, thereby object segmentation is automated or object matching becomes precise in pixel level. In this talk, I�ll introduce two of my recent works in this direction. Specifically, I�ll introduce how object matching model can be integrated with two different types of segmentation models, namely the Markov random field model and the active contour model, to achieve automatic object segmentation. Discussions on their uses in radiographic image analysis would be encouraged.

Biography :

Email: wox.jywang@gmail.com