A_Study_On_Binary_Image_Representations

A study of various transformations on compressed binary image representations

Setup

All implementations are in Python 3.6.5. An Anaconda setup in Linux is preferred, however a Windows setup would also do.

• Anaconda for Linux (Python 3.6) https://docs.anaconda.com/anaconda/install/linux/

• Anaconda for Windows (Python 3.6): https://www.anaconda.com/download/

Install external dependencies as:

pip install numpy Pillow opencv-python

Algorithms and their usage

The following are the three areas that have been studied as a part of this work:

1. Inter-conversion between run-forests and quadtrees:

   • For run-forest to quadtree conversion, run

      python run_forest_to_quadtree.py path_to_run_forest_file
     

   • For quadtree to run-forest conversion, run

      python quadtree_to_run_forest.py path_to_file_containing_quadtree_codes
     

2. Geometric transformations on run-forests - The following geometric transformations have been implemented:

   • Scaling
     • Subsampling
     • Zooming
   • Rotation - A single module contains two different functions for clockwise and anticlockwise rotations.
   • Translation - A single module contains two different functions for horizontal and vertical translations. The scripts can be run as:

      python run_forest_geo_op path_to_run_forest_file
     

     where geo_op is the name of the geometric operation which has to be performed. Eg:

      python run_forest_translation.py img_run_forest.txt
     

     The main functions implementing the geometric operations have the same names as the respective operations they implement and they take as arguments a run-forest and a factor, if the operation requires one.

3. Morphological transformations on run-forests - The following morphological transformations have been implemented:

   • Erosion
   • Dilation
   • Hit-or-miss transform The scripts can be run as:

      python run_forest_morph_op path_to_run_forest_file
     

     where morph_op is the name of the morphological operation which has to be performed. Eg:

      python run_forest_erosion.py img_run_forest.txt
     

     The main functions implementing the morphological transformations have the same names as the respective transformations they implement and they take as arguments a run-forest and a structuring element, using which the run-forest is transformed.