Wednesday, September 30, 2015

Average Image Art creations (AIAs).

I have created a facebook page where I share my collection of Average Image Art (AIA) pieces.
The Page is here:
By using different images of similar subject, and aligning the same features to make that feature sharp and other features more abstracted we get really interesting results that look different and unique from original input images.
Here's an AIA piece of averaging 7 different images of Pharaoh Tutankhanmun.

The result is sometimes very abstract if the source images used are very different in perspective.
Or if the source images are similar you something like above where the eyes are very sharp and in focus. and everything else has abstract looking transparencies.
Here's an a very abstract example of a red car (also from 7 different images of red cars at similar but slightly different perspectives, the center of front and back wheels were used to align).
I hope you enjoy these pieces, and maybe even create your own to share on my facebook page.

You can also see more of these AIAs (Average Image Arts) here:
If you compose some AIAs of your own, please share on my facebook page here: Average-Image-Art-AIA Facebook Page

Friday, September 25, 2015


This is the interactive version of this link: Your Super Villian name and Title
The interactive version is here: Interactive Program to get your Super Villian Name and Title
The interactive version is just like the image shown in 1st link above but it's a little program that will prompt you for your first name, last name, day of month you were born and middle name (if you have one) if not just enter blank.  And it'll tell you what your Super Villian Name and Title is.

Thursday, September 17, 2015

Neural Network Generated Artwork.

To generate your own Neural Network Generated Artwork.
You need 2 images.
One is your original art work, and the other is the style you want to copy.
You upload it to
and then wait for them to send you an email when it's done thinking/processing/painting.

What's really interesting is that if you use the same input images and repeat the process, you won't get the same result.  Just like how us humans paint, we can't paint an image identically every time like a printer, the neural network generated artwork is slightly different each time you submit.

Below is an animation of 7 frames Neural Network Generated Artwork where same input images have been used but the 7 frames are slightly different but generally still represent the same painting.

Sunday, September 13, 2015

How to make my own deep dream images/video

to make your own deep dream images, you can just go to

To make video(s) of deep dream images like a video where it keeps zooming in, you can keep feeding your image and use the result as input again and feed it through again and zoom/scale up each time and generate frames and put them all together using windows movie maker.

Or if you don't want to sit there and manually upload to, you can use a script i wrote to automatically submit to and scale up your image at each zoom step.

The script is python and requires you to install some additional packages for python and imageMagick (used for the scaling/zooming step).

Detailed instructions on how to use the script and what to install can be found here: