Image Transformation and Restoration using AI

Transformation : A marked change, as in appearance or character, usually for the better. (freedictionary.com)

There are many different sites and algorithms that enable transformation of images from one form to another. We discussed the Image2Image technique thatat the beginning of this Chapter also enables us to create variations or versions of one image into several alternative renderings. I have used this method very early on with Dall-e-2 and later with NightCafe when I was learning the basics of AI Image generation.

Restore: the act or process of returning something to its earlier good condition or position.(Cambridge Dictionary)

Image restoration alludes to the process of enhancing or improving the quality an image which has experienced some sort of defect or damage. Too much noise, blurriness, and missing sections of the image could be some of the typical defects in that picture. AI algorithms can map out the patterns or the features in an image, and use some predictive or corrective techniques to correct the distortions, blurriness,and much more.

Below is an image of I had generated on Nightcafe, using the prompt "Picture of a woman and a Cow in rural India". With some other keywords such as season, colours, painting style and so on. Using the inbuilt tool in Nightcafe, I generated two more variations of the same image, and the results of the same are presented below. Also see a short video generated by Nightcafe captured as the image transformation was taking place.

The below image was generated using Image to Image Transformation.

https://avyas.b-cdn.net/Books/AIBook/ch-ii/Video/Woman-Cow-Image-Transform.mp4

The site Tinywow is one such site that offers many options for editing and transforming images and other file types. Below are a couple of screenshots form the Tinywow website, that show the changes or transformations that can be made to images. Some of these transformations are tagged or classified as "AI driven" transformations.