| 摘要 |
A method and system is presented for automatically selecting a region of interest (ROI) within an image of an object, for example when performing non-rigid image registration between the image and another image of the object. In this way, the need for user interaction can be minimized or eliminated during image registration. The ROI is determined by defining an entropy measure H of the image, and selecting the region within the image in which the entropy measure is maximized, In this way, the ROI is optimized to contain as much information as possible. In one embodiment, the entropy measure H is a modified Shannon entropy, defined by H=-SigmaI P(I) log P(I), where I is the value of the image intensity level, and P(I) is the probability of an image intensity value I occurring within the ROI. |
