VisionX V4VCONNECTVisionX V4
NAME

vconnect - compute connected segment descriptors

SYNOPSIS

vconnect [if=<infile>] [of=<outfile>] [th=<thresh>] [of2=<dfile>] [f=<first>] [l=<last>] [-2d] [-a] [-z] [-c]

DESCRIPTION

Vconnect generates descriptors for each of the connected segments that pixel values greater than or equal to a threshold for a byte image or image sequence. That is it performs connected component labeling and generates a descriptor for each unique connected component. For an image sequence 3D connected regions are detected by default. A 3D graph file is generated which contains for each segment a unique label, the location of the center of gravity, and optionally a descriptor vector. The descriptor vector contains the bounding box of the segment and its low order moments.

The format of the optional 14 element feature vector is as follows: (0) feature vector key , (1-6) region bounding box , moments m000 (7), m100 (8), m010 (9), m001 (10), m111 (11), object id (12), and internal region id (13). The feature vector key has the value 2 when the -a flag is specified; otherwise it has the value 1.

CONSTRAINTS

Only unsigned byte and binary (bit) images are currently accepted for input.

OPTIONS

if=

Specifies the input file

of=

Specifies the name of the output file.

-2d

Extract 2D connected regions only; i.e. do no allow connections between frames.

-a

The low order moments of a segment are based on a uniform density for each pixel of 1. By default the density of a pixel is its grey level value.

-f

The -f option specifies that feature descriptors are to be generated; by default only the label and location of the segments are output.

-c

The -c option corrects for missing scan lines in (interlaced) video data, this will assume that pixels between pixels that are above threshold are themselves above threshold.

th=<thresh>

Specifies the threshold for segmentation to the value <threshold>. The default value for <threshold> is 1.

f=<first frame>

Specifies the first frame of the image set to segment. Defaults to the first frame.

l=<last frame>

Specifies the last frame of the image set to segment. Defaults to the last frame.

of2=<filename>

Specifies an optional debug output file that contains images with incomplete label values.

AUTHOR

A. P. Reeves and B.P. Kelley