Figure 7
From: Ultrasonic actuation of a fine-needle improves biopsy yield
General description of the classification process. During the training phase a number N of randomly selected unclassified objects objects are selected from the data and firstly assigned to one of the output classes (tissue fragment, cell cluster, single cell, debris) by a user. Several morphological features (area, eccentricity, circularity, etc.) are then extracted from such objects and used as classification criteria during the classification phase.
