Pixit

This is a simple image editor to fill a particular need. It currently has two mode of editing: individual pixel and selected region of interest (defined as ROI). The basic workflow is:

• select a pixel or a region of interest
• replace it with a new RGB value
• save the image

What not OpenCV? OpenCV has a function call selectROI() which allow you to select a rectangle of region of interest as well. There are several problems with it:

• Multiple ROI operation doesn't seem to work
• Have no support single pixel operation
• Have no support of polygon ROI

Thus, this simple DIY script. Please report bugs/requests to issues.

Set it up

This utility has two dependencies: Python 3 and OpenCV (3.x). The version dependency is not critical: it can work under Python 2 with a bit tweak. Here is one way to set it up using Anaconda:

• Install anaconda
• Install OpenCV: conda install opencv
• Activate either the base or whatever the new environment you desire.

Operations

To run it:

python pixel.py rabbit.jpg

You need to switch between image window and console window for most operations:

Select ROI	Replace with new value

• To quit: ESC key on the image window
• 'p' for selecting a point.
• 'r' for selecting a rectangular area of interest.
• 'P' for selecting a polygon, right click to end a selection.
• 'c' for canceling previous selection.
• Once a selection is made and ready to replace, on the image window, key in R, then go back to console window to type in new RGB value, such as 255 0 0 and Enter to confirm.
• To save the image, key in s or S on the image window, console window will display the new image name, which is composed by old image base name, suffixed by -new.