Architecture

This page describes the architecture of the software.

It helps modifying and maintaining the software. It also demonstrates how to combine numpy (or scipy) with gimp (gimpfu_numpy_converter.py).

Goal

Create a tool to generate anaglyphs from (scanned) flat images arranged in depth.

Constraints

The selected solution has the following constraints: Plugins must be written in python or scheme.

Context

The context of the plugin is gimp (who owns/calls the plugin) and additional required libraries for signal/image processing.

Solution

• Real 3D renderer (blender, povray) were considered to be too complex to use for this task.
• A gimp plugin seemed to be a possible solution.
  • It allows easy composition of scanned images.
  • It also allows to scan and pre process the images.

Building Block View

The context of the plugin is gimp (who owns/calls the plugin) and additional required libraries. The plugin itself is organized as follows:

• Module create_anaglyph_plugin.py (component "Plugin"): The main plugin code (control code)
• Module anaglyph_algorithm.py (component "anaglyph_algorithm", class anaglyph_algorithm.Anaglyph): the core algorithm
• Module depth_algorithm.py (component "depth_algorithm", class depth_algorithm.DepthAlgo): functions to determine a layers depth
• Module gimpfu_numpy_converter.py (component "gimpfu_numpy_converter"): functions to convert from/to gimpfu data structures to/from numpy arrays.
  

Runtime View

Here we sketch the activities executed by the plugin (create_anaglyph_plugin.run). The colors match the components corresponding to the described modules above.

Deployment View:

(see Installation)

Concepts

• Data is converted from the gimp format to numpy and back.
• Algorithms process on numpy arrays.