Artifact Images (AI)

The interactive view in all the dimensions and associations does not fare well when exported to two space and the lack of the ability to change the dimensional relationship. It is possible to see the strings that attach the elements to the color vector space. It is also possible to see what objects exist within other objects , such as faces and features of the face which can be selected from those groups easily. It now shows the image as a flat in the back of one of the planes and it uses the elvi to get images from the internet and analyze to identify whether the images are fabricated, retouched or of similar origin. It is odd what shows up in the other dimensions and it becomes so obvious how the images originated and in fact what software was used to create some images. An image that is 1024x1024 in black and white contains a possibility of 2 to the millionth power possibilities. Within that, there are many ways of "looking" at the data and some show traits that one would not suspect would be there and so easily attainable.

One thing that is immediately obvious in images when presented in the vector color view, is that created images are flat vector planes and real images from life or nature are like specific manifolds of vector space. They can be separated with a single look. There is no way for me to present some of the data , particularly those that involve differentials and integrals or combinations that include Fourier. The differentials δn{x,y,z} are the most interesting to me. The second and those which are differentiable to things like the "heavyside" function are also filled with information.



This is the output from my program in 3 dimensions as applied to the color test above It is not difficult to see the nature of and type of the image itself with a simple dimensional view. Here R=X, G=Y, B=Z. The other dimensions of the images are more complex and involve the display of sets and associations in several dimensions and I am having some trouble making a usable interface that allows the kind of actions I require. Specifically I would like to lasso traits and create sets and then rotate the dimensions about those sets , but the best code concept is not coming to me.

In the image above the section of gray scale is obvious as a characteristic. I also see that colors were chosen as 32 levels of color. Once an image is correlated with its spatial and texture aspects it becomes more and more coherent. Real life pictures form distinctly obvious elements that tell me what is in the picture without actually looking at it. Certain shapes and forms come from each type of object. Man made objects like the above are very obvious even in the RGB=XYZ dimensions. I reused code from my puzzle game and did not change the window title and I also did a bad thing and put a "goto" to bypass the puzzle and show this, <*motey hangs his head in shame*>.


atx=((float)(uint8_t)newimage[0])/256.0; aty=((float)(uint8_t)newimage[1])/256.0; atz=((float)(uint8_t)newimage[2])/256.0; glVertex3f(atx, aty, atz);

The four dimensional view is more informative, but that will have to wait until my WebGL plasmoid is done.


This is an image of a living object and I have enabled color as a added dimensions in this view. As a result it has 7 dimensions technically shown at once. I wonder if I could apply that to display other dimensional aspects as color. It is really obvious when the image above is compared to this, that one is organic and one is mechanical.


And the original image with colors for the data points.


And finally a color wheel. I would show the original source, but some of the images are only figments of my computer's imagination and only exist transiently. Clearly it is a constructed form and not natural.

0 comments:

Automated Intelligence

Automated Intelligence
Auftrag der unendlichen LOL katzen