COMPUTATIONAL METHODS FOR THE SEGMENTATION OF IMAGES OF OBJECTS FROM BACKGROUND IN A FLOW IMAGING INSTRUMENT

摘要

In automated computation-based interpretation of images, the accuracy and reliability of the detection of objects, known as "object segmentation," is implemented so as to provide efficient performance. In a multi-step process, objects are first detected (74) and captured into regions of interest (ROIs) (76). Sets of pixels belonging to respective objects are then identified. Preferably object detection is achieved using both a two-dimensional (2D) low pass filter and a 2D edge enhancement filter. Two different gradient based edge enhancement filters are disclosed. One embodiment of the invention defines a (ROI) by first determining the center of objects by executing a plurality of decimations on the filtered image data, and then establishing object boundaries. In a second embodiment the ROI is defined by generating an amplitude histogram of the filtered image data, and for histograms exceeding a threshold determining by pixel which rows are to be included in the ROI.