Skip to main content
EURASIP Journal on Image and Video Processing

Table 1 The classification of clusters

From: Clustered nuclei splitting based on recurrent distance transform in digital pathology images

Cluster class

Criteria

Description

I

Area >average_Area & Eccentricity > 0.8

Generally two overlapping objects, most common

II

Area >average_Area & Eccentricity > 0.6

Three or more objects, with various overlap

III

Area >average_Area & Perimeter > T_Perimeter

Heavily packed objects, with high overlap

IV

Area >3*average_Area

Very big clusters

  1. Where average_Area is estimated for dataset based on nonclustered objects of interest; T_Perimeter is the perimeter of the circle with average_Area. Examples are presented in Fig. 5
Back to article page