EE/CSE 576: Image Understanding

Spring 1999

Course Software: ipeg output format

The output file produced by ipeg looks like the output of fex and lpeg: a number of sections, each starting with text tag and containing zero or more description lines.

ipeg passes through all the objects it is input (the ones produced by fex and the ones produced by lpeg), as well as adding the following groupings:

#Triplet

triples of segments which form barends or Zs.

int	triplet ID
int	segment 1 ID
int	segment 2 ID
int	segment 3 ID
float	start point column
float	start point row
float	end point column
float	end point row
float	junction 1 column
float	junction 1 row
float	junction 2 column
float	junction 2 row
float	total length of segments
float	total length of triplets including gaps
float	quality

#Y_Corner

#TLambda_Corner

triples of segments which form Y or TLambda corners.

int	triplet ID
int	segment 1 ID
int	segment 2 ID
int	segment 3 ID
float	center point column
float	center point row
float	radius
float	quality

#ClosedPolygon

sequences of segments that form closed polygons. The number of fields on description lines of this type is variable.

int	number of segments
int	segment 1 ID
int	...
int	segment n ID
int	number of junctions
float	junction 1 column
float	junction 1 row
float	...
float	...
float	junction n column
float	junction n row
float	quality