A Neural Network Based System for Optimal Medical Image Visualization with Fast Online training Capabilities

摘要

An adaptive hierarchical neural network ased system with online adaptation capabilities is develooped to automatically adjust the display window width and center for medical images. Our windowing system possesses the online training capabilities that make the adaptation of the optimal display parameters to personal preference as well as different viwing conditions possible. in addition, our system can easily adapt to new types of medical images simply by including these images into the training set. The online adaptation capabilities are achieved through the use of the hierarchical neural networks and the development of new width/center ampping algorithm. The large training image set is hierarchically organized for efficient user interaction and effective re-mapping of the width/center settings in the training data sset. The hierarchical neural networks are then re-trained for the new training data set after the mapping process or after the inclusion of new images to adapt its window width/center estimation to the user's preference and the viewing coditions.