独立成分分析を利用した多値ボリュームデータの対話的セグメンテーションと可視化に関する研究

抄録

An interactive segmentation and visualization technique for volume data is presented which uses independent component analysis (ICA) and a three-dimensional histogram of the volume data. Volume segmentation becomes equivalent to clipping of the three-dimensional histogram in ICA space and results in a transfer function for hardware-accelerated rendering which provides immediate visual feedbacks to facilitate further exploration of the data. By applying this technique to a variety of volume data sets it is demonstrated that in many cases volume segmentation in ICA space is more intuitive and effective than in the original voxel space.