An Unsupervised Point Alignment Detection Algorithm
Please cite the reference article if you publish results obtained with this online demo.

This is an implementation of an Point Alignment Detector.

Use an example:




Border and line




Masking by Cluster

100 noise points

1000 noise points

Alignment and noise

Perfect Alignment

Simple Alignment

Masking by Structure

Draw Points

  • Click "Draw Points" to start drawing points.
  • Optionally add random noise.
  • Once you have finished, click "Run" to run the detectors.

Upload Data

Upload your own dots files to use as the algorithm input.


The uploaded text file shall be formatted such that each line contains a pair of x, y coordinates, separated by spaces or tabs. Coordinates must be real numbers between 0 and 1 (it is assumed a unit square domain). 1000 points is the maximum allowed.
0.1523418 0.8593640
0.1933569 0.8242082
0.2460906 0.7851462