Connected Components Analysis

Connected components analysis (CCA) is a method to identify and label all the connected components or distinct objects in an image, where a connected component is a set of pixels that are connected via some predefined criterion. The criterion used in Dragonfly for defining connected components is connectivity, in which pixels are considered connected if they share a common face, edge, or corner (see Connectivity).

The output of a connected component analysis is a labeled multi-ROI. A different color is assigned to each label, making it easier to identify and analyze the connected components in the multi-ROI. Subsequently computing measurements for each connected component, such as size, surface area, shape, center of mass, and other parameters, provides the opportunity to quantitatively analyze objects and classify them with extracted features and properties. Refer to the topic Measurements and Scalar Data for Multi-ROIs for information about the available measurements and metrics for multi-ROI. Refer to the topics is Analyzing and Classifying Measurements for information about performing a cross-table analysis of feature vectors.

You can start a connected components analysis from the contextual menu that is available for regions of interest and multi-ROIs as follows:

New Multi-ROI (6-Connected)… Automatically creates a new multi-ROI, in which each group of connected voxels is labeled as a distinct object. Propagation is done by strictly using the 6 faces adjacent to the current seed and will result in the minimum number of connected pixels (see 6-Connected and 26-Connected).

New Multi-ROI (26-Connected)… Automatically creates a new multi-ROI, in which each group of connected voxels is labeled as a distinct object. Propagation is done by strictly using the 6 faces, 12 edges, and 8 corners adjacent to the current seed and will result in the maximum number of connected pixels (see 6-Connected and 26-Connected).