A trap...

It's a trap... Messing around with black and white images of my very own venus fly traps using the DeOldify software - DeOldify

DeOldify is a software tool that can be used to colourize black & white images made possible through artificial intelligence and deep learning. It is free and open source to use so a very good resource to have on hand. As part of the project they have set up Google Notebooks which allow users to access and process their own images.  Above is a colourized version of my photo using a render factor of 14. Below is the original photo I took of the trap prior to processing in Photoshop.

This exercise was completed as a test. It's curious to see how the colours differ from the original prior to processing. 

The default value for rendering images in the program is set to 35. They say that this value has "... been carefully chosen and should work -ok- for most scenarios (but probably won't be the -best-)." To explore in more detail how it effects the output, I used a black & white image and set the render factor to different numbers incrementally to monitor how it affected the image. I was curious to see how the AI program interpreted the "old" (not really) desaturated image in terms of colour versus my original colour image. I guess I wanted more insight into the type of perspective the AI has when it comes to "seeing". 

The render factor determines the resolution at which the colour of the chosen image is rendered: 
- Lower render factor will process faster with the possibility of more vibrant colours. 
- Higher render factor can be better for higher quality images but may result in less vibrant colours.

For example, they state that older and lower quality images usually benefit by lowering the render factor.