| 摘要 |
The present invention may include segmenting an image at a first resolution level and a second resolution level, wherein one or more parameters of a segmentation algorithm are trainable via user classification feedback, extracting features from a first plurality of segment primitives and a second plurality of segment primitives, wherein one or more parameters of a segment feature extraction algorithm are trainable via user classification feedback, building a first and second segmentation hierarchy by generating one or more clusters of the first plurality of segment primitives and the second plurality of segment primitives, extracting one or more features from the first segmentation hierarchy and the second segmentation hierarchy utilizing a hierarchy feature extraction algorithm, determining an inter-level relationship between the clusters generated for the first resolution level and the second level, and automatically classifying one or more tissue elements of the tissue specimen via a user-trained classification algorithm. |
