Add more image processing methods

[ankitaggarwal011]

The different image processing techniques with cellular neural networks can be realized by changing the coefficients of the control and feedback templates.

Significant research had been done for determining the templates needed for realizing various image processing functions. One can refer to this published template library for numerous examples.

Adding these new functions would require adding new template coefficients corresponding to these functions. This would be exactly similar to current functions such as edge detection except the coefficients would be different.

Any suggestions are welcome. Thanks!

[bryceandpeas]

Hi, how is the t variable for each technique determined? I understand that the start value will always be 0, how are the stop and number of samples values determined? I'm struggling to connect it to the document you linked to (whereas everything else is pretty self-explanatory).

[ankitaggarwal011]

@brycefury, Thank you for your interest. This is an excellent catch, somehow I missed explaining the role of variable t.

The variable t refers to the time component while performing the integration. The line L81 defines t = np.linspace(0, 10.0, num=2), in which the start = 0, end = 10.0 and number of sample (num) = 2. Normally, it doesn't take more than the specified number of time points for the output to converge, however, this value is configurable as some type of processing might need more time points.

If you will display the output with each time point, you can actually see the evolution and the final convergence to the output, it actually looks pretty cool. We can add this feature of observing saving each time point output in the library if your interested.

I'll add this information as a comment in the code and as a note in the documentation. Thanks for bringing this up. Please let me know if you face any other problem while adding new image processing methods or have more questions and thanks for your contribution.

[ankitaggarwal011]

Added a new method of Optimal Edge Detection via #38, #39, #40.