TECHNOLOGY | Identifying demon-people assailants via chroma-facial signatures

What I knew two years ago...

Long before I knew as much about chroma as I do now (over two years ago), the most I could say about it is that it tended to concentrate in people around the face, which had the benefit of revealing the inner demon of someone possessed [see PHOTOS | Underlying demon revealed in video in motion], and of producing illusory masks of a kind for those with the know-how [see VIDEOS | Demonic chroma masking (real-life Halloween masks)], but that it was otherwise problematic, in that it saturated digital images with excessive color noise.

...what I know now

Even though my understanding of chroma hasn't increased much in those two years later, I've learned to make prodigious use of it when it's present with a digital camera and GIMP or Photoshop, even though chroma is the culprit in every bad digital image made. For example, since chroma lights up demons, so to speak, I can capture them with my iPhone when they are cloaked and in the dark after applying basic scientific imaging techniques to the images; I've also developed a way to detect demonic activity through the analysis of digital images that were made in the presence of chroma.

Newly discovered use for chroma

This post adds yet another way chroma can be used to your advantage during periods of high demonic activity, specifically, face-tagging demon-people assailants, whose use of demonic weapons and possession by and/or interaction with demons expose them to high levels of chroma, which concentrates around the face—just like I noted two years ago—allowing for the creation of a facial signature from the assailant's face from far away and in the dark, even when the source image is shows no distinguishable facial features at all. What's more is, all of this can be done with just an iPhone and GIMP or Photoshop.

A facial signature, by the way, is a graphical representation of the imprint a chroma-saturated face makes on a digital image sensor. They can be obtained from very far away, and while they don't look exactly like the person—rather, more like a caricature—the signature will always be unique to that person, and consistent across multiple images. Once you've matched a facial signature to the person's actual face, you can identify them using off-the-shelf, readily available consumer electronics and free image-processing software, even when your own eyes can't make out who they are.

Following is an example of the facial signature a demon person, who, in the early morning hours, launched a demon into my building in order to attack (at, of course, the behest of the Voices Demons, and, of course, during an anger management ritual):

The facial signature of a demon-person assailant (below) looks like a cartoon version of a drunken souse

Not impressed? Then wait until you see what this was made from:

Source video from which the demon-person assailant's facial signature was derived (above)

This is the portion of a surveillance video in which the assailant is shown guiltily looking up at my apartment as he slinks past my bathroom window on the second floor in the wee early morning hours, and from which the facial signature shown above was derived:

Location of the iPhone making the video (second-floor, rear of building), and the location of the assailant (sidewalk)

It is not his face that is reconstructed in GIMP; it's the image data created by the iPhone image sensor as imprinted by the EMF radiation emitted by the culprit. The imprint is influenced by the man's face, but is not a perfect replica of it, and is not related in any way to normal light that would reflect off his face, if there were enough of it.

In other words, there are two faces reflected from this man: one made of reflected EMF radiation; the other, natural light. What's interesting to note is that the lack of normal light sufficient to illuminate his face is what allowed decent imprint of the EMF radiation; that's because it's weaker than normal light, which, if stronger, would have washed out the facial signature, and revealed only his natural face:

The dopey-eyed boozer in a fishing hat on the left and the shadowy assailant on the right are the same demon person

So, in essence, chroma allows for identification of demon-person assailants in the dead of night, from far away, in the dark, which is a real boon for anyone interested in knowing who's working as an agent of the anti-Christ, and carrying and deploying invisible deadly weapons.

Chroma image-mapping by filtering natural light
In both GIMP and Photoshop, you can filter out natural light, leaving only the image data left by chroma (or the EMF radiation therefrom) by blending (or overlaying) the difference of differences between two consecutive video still frames.

Here's what that looks like in GIMP, where frame1 and frame2 are the two consecutive still frames:

Blending the difference of differences between two consecutive video still frames filters natural light (shown in GIMP)

Here's what that looks like in Photoshop:

The blending modes and stacking order of layers made from two consecutive still frames for filtering natural light using Adobe Photoshop

With one exception, the procedure for visually isolating chroma from natural light between two stills frames is the same—the exception being how layers are combined to produce a single composite layer. In GIMP, two (or more) layers can be combined by creating a new-from-visible layer; in Photoshop, two (or more) layer copies must be merged, and, if desired, the original layers either disabled or deleted.

With those differences in mind, here are step-by-step instructions for producing an image like the one shown above in GIMP (Photoshop instructions available upon request):

Create a new image with two layers using two adjacent video still frames, chronologically
Duplicate both layers
Invert both duplicates
Set the blending mode of the duplicate layers to Difference
Create a new-from-visible layer
Set the blending mode of the new-from-visible layer to Difference
Disable both the original and copy of the second still frame-layer
Create another new-from-visible layer
Stack a copy of the second still frame-layer over the new-from-visible layer
Set the blending mode of the copied layer to Overlay
Flatten the image

The layer stack should look identical to the one shown above, and the results approximately the same. The degree of color noise from the chroma imprint on the digital sensor will vary, but even the slightest shift in position by the subject between the two frames should produce very noticeable results.

Detecting motion of invisible demons, other entities in video

Following is a GIMP Python-Fu script that applies the procedure above to each still frame in a video, which enables detection of cloaked (semi-invisible) demons, demon people and objects in transit in front of the camera. While the process may not render them readily distinguishable as such, their general shape, size and path will be easily discernible:

from gimpfu import *
from os import listdir
from os.path import isfile, join

def list():

mypath = "/Users/jamesalanbush/Desktop/test/"
    onlyimages = [ f for f in listdir(mypath) if isfile(join(mypath,f)) ]
    onlylayers = [ f for f in listdir(mypath) if isfile(join(mypath,f)) ]
    del onlyimages[-1]
    del onlylayers[0]
    
    for i, j in zip(onlyimages, onlylayers):
        try:
            if i.lower().split(".")[-1] in ("png", "jpg"):
                img = pdb.file_png_load(mypath + i, mypath + i)
                disp = pdb.gimp_display_new(img)
                
                # Duplicate frame1, invert and Difference
                frame1_layer = pdb.gimp_image_set_active_layer(img, img.layers[0])
                frame1_layer_copy = pdb.gimp_layer_new_from_drawable(pdb.gimp_image_get_active_drawable(img), img)
                pdb.gimp_layer_set_mode(frame1_layer_copy, DIFFERENCE_MODE)
                pdb.gimp_image_add_layer(img, frame1_layer_copy, 0)
                pdb.gimp_invert(frame1_layer_copy)
                
                image_id, layer_ids = pdb.gimp_file_load_layers(img, mypath + j)
                for id in layer_ids:
                    new_layer = gimp.Item.from_id(id)
                    pdb.gimp_image_add_layer(img, new_layer, 0)
               
                # Duplicate frame2, invert and Difference
                frame2_layer = pdb.gimp_image_set_active_layer(img, img.layers[0])
                frame2_layer_copy = pdb.gimp_layer_new_from_drawable(pdb.gimp_image_get_active_drawable(img), img)
                pdb.gimp_layer_set_mode(frame2_layer_copy, DIFFERENCE_MODE)
                pdb.gimp_image_add_layer(img, frame2_layer_copy, 0)
                pdb.gimp_invert(frame2_layer_copy)
                
             # Create a new-from-visible with frame 2/Inverted
                nfv_layer_diff = pdb.gimp_layer_new_from_visible(img, img, "Diff: Frame 2/Inverted")
                pdb.gimp_layer_set_mode(nfv_layer_diff, DIFFERENCE_MODE)
                pdb.gimp_image_add_layer(img, nfv_layer_diff, 0)
                
                # Disable frame2
                pdb.gimp_item_set_visible(img.layers[2], FALSE)
                
                # Reorder frame2 inverted and set to Overlay
                pdb.gimp_image_raise_layer_to_top(img, frame2_layer_copy)
                pdb.gimp_layer_set_mode(frame2_layer_copy, OVERLAY_MODE)
                
                pdb.file_png_save_defaults(img, pdb.gimp_image_flatten(img), pdb.gimp_image_get_filename(img), pdb.gimp_image_get_filename(img))
                
                pdb.gimp_display_delete(disp)
                pdb.gimp_displays_flush()
              
        except Exception, error:
            print error

This script processes a PNG-file sequence made from a video using Adobe Media Encoder, placed in a folder named test. Simply change the path variable to the directory containing your PNG or JPG file sequence.

There are scores of such videos in my collection, waiting to be processed by this script; they will be posted as soon as they are ready, so be sure to check back. In the meantime, here's part of one, made from the video used in the example in this post:

NOTE | I just now finished the script [5p, 12-21-2014], so the wait will not be much longer.

Tools for building videos from image sequences
An alternative to Adobe Media Encoder CC for building video files from a series of processed images is the open-source ffmpeg software package; with it you can convert your video files into a sequence of PNG files for processing by the GIMP script:

Using ffmpeg to convert a video file made on my iPhone into a sequence of PNG files for processing by the GIMP script posted above

After you've run the GIMP script on the PNG files, you can reconstruct the video file using mencoder:

Using MPlayer to encode a video file from a sequence of processed images

ffmpeg and mencoder are fast, efficient, and worked when Media Encoder would not.

NOTE | Soon, the GIMP script will be updated to use ffmpeg to decode video files into a sequence of individual images, process them in GIMP, and then reconstruct a video file using mencoder all in one step.

The GIMP Plug-in
GIMP users can install a plug-in that performs the mask-light/isolate-chroma procedure shown in the example video above:

My plug-in can be used in other GIMP scripts by calling python-fu-list

It installs a Process menu, and is called Mask Light:

Click Mask Light on the Process menu to use the GIMP plug-in

You can download it from Google Drive; to install it, place it in /opt/local/lib/gimp/2.0/plug-ins/ or equivalent, and grant executable permissions (i.e., chmod +x list.py).

Coming up: image filters to improve nighttime video footage
To enhance facial-signature capture, image filters for improving nighttime video footage is in the works. This still frame taken from the video used in this post shows the results so far:


Before	After

Other than a little sharper, improved contrast and somewhat brighter, the difference doesn't look like it amounts to much; however, adjustments made to dark footage must be carefully calculated so that any change in light levels does not cause dithering, banding, or out-of-gamut distortion.

Here's another iteration of the night-time video enhancement script:


Before	After

The Life of a Demoniac

Search This Blog

TECHNOLOGY | Identifying demon-people assailants via chroma-facial signatures