Installation and usage

Required software

Following tools needs to be installed:

• Python 3.x
• pillow (Python image library)
• numpy (Python numeric operation library).

On Ubuntu for Windows it can be done this way:

apt install python3-pip
pip3 install pillow
pip3 install numpy

Configuration (optional)

Configurable parameters are stored in srconfig.py file.

Usage

Run:

python3 automate.py (filename of image in ./samples directory)

Example

python3 automate.py circuit.tif

This will create low resolution samples in ./input directory and then perform high resolution restoration based on them.

Example of output:

Testing algorithm for circuit.tif, iterations: 5, scaling factor: 2.000000
Creating Camera Model
Samples created cuccesfully, restoring (this might take time)
Estimate Motion Between Sample And Original Image
S_0_0.tif
S_0_1.tif
S_0_1.tif: (0, 1)
S_0_2.tif
S_0_2.tif: (0, 3)
S_1_0.tif
S_1_0.tif: (1, 0)
S_1_1.tif
S_1_1.tif: (1, 1)
S_1_2.tif
S_1_2.tif: (1, 3)
S_2_0.tif
S_2_0.tif: (2, 0)
S_2_1.tif
S_2_1.tif: (2, 1)
S_2_2.tif
S_2_2.tif: (2, 2)
samples loaded
Creating Camera Model
Size Of Estimated Original: 84x104
 0: estimation error: 90.053265
 1: estimation error: 67.434944
 2: estimation error: 52.375802
 3: estimation error: 40.919872
 4: estimation error: 32.288271
Total elapsed time: 0.27477276722590127 mins

Note that errors of estimation decreases with each iteration. Output image is in ./samples named super_resolution.tif