VapourSynth-cv_ccl_filter

Vapoursynth Connected Components Label Filtering from OpenCV.

Description

Using OpenCV's connectedComponentsWithStats to compute the connected components labeled image of a boolean image and also get statistics for each label. Then filter the labeled content based on the statistical results.

Require VS API >= 4.0

Usage

Exclude the labels above ccl_thr

core.cv_ccl.ExcludeCCLAbove(clip mask, int cc_thr[, int connectivity, int ccl_type, int cc_stat_type])

Exclude the labels under ccl_thr

core.cv_ccl.ExcludeCCLUnder(clip mask, int cc_thr[, int connectivity, int ccl_type, int cc_stat_type])

Read all the stats from cv::connectedComponentsWithStats to FrameProps

core.cv_ccl.GetCCLStats(clip mask[, int connectivity, int ccl_type])

For example, if you want to remove all the connected components smaller than 2500 pixels (in area) inside your mask clip, you can simply write:

core.cv_ccl.ExcludeCCLUnder(mask, 2500)

Sample

Original

ExcludeCCLUnder 2500

ExcludeCCLAbove 2500

If you want to get all the ccl stats data, using core.cv_ccl.GetCCLStats, you will get FrameProps like:

_CCLStatNumLabels
_CCLStatAreas
_CCLStatLefts
_CCLStatTops
_CCLStatWidths
_CCLStatHeights
_CCLStatCentroids_x
_CCLStatCentroids_y

which should contains array (except _CCLStatNumLabels) if the input clip isn't a blank clip. Note that the first label is always the background label, so you will like to get the frame prop like this

f.props.get('_CCLStatAreas', None)[1:]

Parameters:

• mask

  Clip to process. Binarize Gray8 clip only.

• cc_thr

  Use to compare with the results of statistics with cc_stat_type.

• connectivity

  8 or 4 for 8-way or 4-way connectivity respectively. Exact param in cv::connectedComponentsWithStats.
  Default 8

• ccl_type

  cv::ConnectedComponentsAlgorithmsTypes
  connected components algorithm

  • 0 - SAUF algorithm
  • 1 - BBDT algorithm
  • 2 - Spaghetti algorithm

  Default -1 Auto (Spaghetti algorithm)

• cc_stat_type

  cv::ConnectedComponentsTypes
  connected components statistics

  • 0 - The leftmost (x) coordinate which is the inclusive start of the bounding box in the horizontal direction.
  • 1 - The topmost (y) coordinate which is the inclusive start of the bounding box in the vertical direction.
  • 2 - The horizontal size of the bounding box.
  • 3 - The vertical size of the bounding box.
  • 4 - The total area (in pixels) of the connected component.

  Default 4 Area.