Prayer Man is a Man

[Andrej Stancak]

While it seems that there is no dispute on this Forum about the impossibility of the alleged female to represent the full figure of Prayer Man, it  may still be useful to point out that Prayer Man having the head aligned with the head of the alleged female would not match Prayer Man's figure in Darnell. The person reaching only to the "head" of the trimmed image would be either much shorter than Prayer Man in Darnell or stand on the third top step (Prayer Man in Darnell effectively stood on the second top step albeit with only his left foot).

Here is an approximate alignment of Wigeman's and Darnell's doorway with guiding lines. If the top of Prayer Man's figure in Wiegman would only reach to the top of the alleged female (the trimmed image), the distance from the horizontal bar in the aluminium door frame would be much larger than in Darnell. In contrast, the top of the head for the full figure in Wiegman aligns with the top of Prayer Man's head in Darnell pretty well given slightly different azimuth and elevation angles in both pictures. The purple line connects the right hand in both Prayer Man figures.

Using the figure of Buell Wesley Frazier as a measuring stick, the top of Prayer Man' head reaches to the lower aspect of Frazier's chin in Darnell and if the full figure of Prayer Man is used in Wiegman. In contrast, the top of the head of the alleged lady would only reach to the level of lower ribs on Frazier's chest.

I hope this ends the odd story of alleged female face in the Wiegman still.

 

 

 

Edited August 27, 2020 by Andrej Stancak

[Sandy Larsen]

48 minutes ago, Andrej Stancak said:

https://thejfktruthmatters.files.wordpress.com/2020/08/pm_arrowsl.jpg

 

I hope this ends the odd story of alleged female face in the Wiegman still.

 

 

Andrej's image shows that the so-called Prayer Woman's face is actually the chin, neck, and upper-chest of Prayer Man.

Good work Andrej. Wish you had been here when this was being debated.

 

Edited August 27, 2020 by Sandy Larsen

[Andrej Stancak]

2 hours ago, Mathias Baumann said:

Andrej,

You could be right about that. Maybe PULSE could be used to identify Prayer Man by trial-and-error, that means by feeding it different images of different people and see which output looks most realistic. I don't know.

But I see a technical problem here:

So it appears you need some special hardware. CUDA was developed by Nvidia. I have a AMD RADEON HD 6450 installed in my computer. I don't think it'll support CUDA, but I'll give it a try anyway.

 

I see the problem. They seem to run the program using the graphic card processor. My laptop also has an AMD RADEON card and, threrefore, I cannot help at the moment. I may have access to a proper computer hosting an Nvidia card at some point in the future. The article also mentions alternatives to AI, maybe those could be easier to implement.

[John Butler]

3 hours ago, Andrej Stancak said:

While it seems that there is no dispute on this Forum about the impossibility of the alleged female to represent the full figure of Prayer Man, it  may still be useful to point out that Prayer Man having the head aligned with the head of the alleged female would not match Prayer Man's figure in Darnell.

Are you saying this is an altered scene?

[Andrej Stancak]

2 hours ago, John Butler said:

Are you saying this is an altered scene?

John, not at all, I believe both scenes, the Darnell and the Wiegman, are genuine. My composite picture was meant to say that if Prayer Man's head would be the alleged female face which some people claim to be Prayer Man's head, Prayer Man would be a dwarf, only reaching to the chest level on Buell Wesley Frazier's figure. 

[Mathias Baumann]

10 hours ago, Andrej Stancak said:

I see the problem. They seem to run the program using the graphic card processor. My laptop also has an AMD RADEON card and, threrefore, I cannot help at the moment. I may have access to a proper computer hosting an Nvidia card at some point in the future. The article also mentions alternatives to AI, maybe those could be easier to implement.

I wish you good luck with that and am curious to see the results.

Now I realize this is off-topic, but there's something else I'd like to hear your opinion on, being an expert in photography.

Please look at these three photos of David Morales:

What do you think is the dark, elongated area above his left eyebrow?

[Andrej Stancak]

Mathias:

It looks like this man did not have his frontal bone completely convex or flat but had a depression in the middle of the forehead possibly due to a combination of slightly larger brow bossing (ridges) (a very masculine feature) and a comparatively thin frontal sinus that sits between and above the brow ridges. The sunk bone above the browse ridges in the middle of the frontal bone causes two more ridges on each side - above the brow ridges and below the hairline. It look natural to me. This man could have also sustained an injury to the frontal bone causing this depression in the middle of the forehead.

Edited August 28, 2020 by Andrej Stancak

[Sandy Larsen]

Some photos of Morales reveal a heavy scar through the eyebrow of the same eye. Makes me wonder if an accident caused both the eyebrow scar and the that thing above his eyebrow, i.e. another scar.

https://alchetron.com/cdn/david-snchez-morales-c03f597d-4c43-46a7-8bed-5f6d42852cb-resize-750.jpeg

 

[Mathias Baumann]

6 hours ago, Andrej Stancak said:

Mathias:

It looks like this man did not have his frontal bone completely convex or flat but had a depression in the middle of the forehead possibly due to a combination of slightly larger brow bossing (ridges) (a very masculine feature) and a comparatively thin frontal sinus that sits between and above the brow ridges. The sunk bone above the browse ridges in the middle of the frontal bone causes two more ridges on each side - above the brow ridges and below the hairline. It look natural to me. This man could have also sustained an injury to the frontal bone causing this depression in the middle of the forehead.

 

Andrej,

thank your for your opinion. It is my belief that Morales indeed had a scar. I think he could have been the man observing Oswald when he was handing out leaflets in New Orleans. Jim Garrison describes him as a muscular Latino and also mentions the scar. Morales was indeed quite chunky and also half-Mexican.

[Andrej Stancak]

Returning to the possibility of increasing the resolution of existing Prayer Man images, a more generic approach appears to be deep learning for single-image super-resolution. This has been implemented in Matlab environment: 

https://www.mathworks.com/help/deeplearning/ug/single-image-super-resolution-using-deep-learning.html?s_tid=answers_rc2-2_p5_MLT

It is a convolution neural network which is first trained on many thousand images how to reconstruct a high-resolution image from a low-resolution version of the same image. Only the luminance channel is used in this method which is good because Darnell's stills are black-and-white images.  The network, after being trained on still images of people, animals and various objects, is able to apply the learned network layers to a sample low-resolution image and increase the high-frequency features (e.g, details of the face) of that image. Comparison of the standard bicubic interpolation and the deep learning method showed an increase in similarity with the original high-resolution image from about 0.25 to 0.87 (1.0 being similarity with its own copy).

This is a major project requiring lengthy computations, some reading and iterative improvements. For instance, using a database of containing only human faces could train the neural network better for the purpose of detecting features in Prayer Man's face than variety of objects (animals, landscapes...). I hope to be able to embark on this in the future, however, this may not be a very near future because of a number of parallel projects and commitments.

 

Edited August 29, 2020 by Andrej Stancak

[Andrej Stancak]

We have authentic handwritten notes taken by FBI agent James Hosty during the first interrogation session with Lee Harvey Oswald on Friday, November 22. The notes were uncovered by Bart Kamp last year and it is the most direct and indisputable evidence of Lee Oswald being in Domino lunchroom on the first floor during the time when motorcade appeared on Dealey Plaza. It is worthy to recall this important discovery again. It says "He went to the 2nd floor to get Coca Cola to eat with lunch and returned to the 1st floor to eat lunch. Then went outside to watch P.Parade."

The question is when did Lee Oswald get to the doorway if he was sitting in Domino room and yet he stood, as some researchers surmise, at the western wall of the Depository doorway in Wiegman film. Here is an attempt to reconstruct Lee Oswald's movements before he appeared in the doorway.

The first question to answer is how much time did it take to a young man to walk swiftly (not running) from Domino room to the western wall of the doorway. Here is a plan of the 1st floor with a trajectory of movements from Domino room to the doorway. The figures next to segments of the route stand for the velocity in that segment. A straight segment in the corridor without obstacles would be walked at a speed 1.79 m/s whereas opening  a door would be at a considerably slower speed (0.45 m/s - tested using stopwatch and different doors). The final passage from the glass door to the western wall was estimated to occur at 0.5 m/s. 

The total distance of the route connecting Domino room with the western wall was 26.2 m and the approximate total walking time was 24.05 ms (rounded to 24 s). 

In the next, I used  Mark Tyler's excellent animation https://www.marktyler.org/mc63.html to identify time stamps of a few events. The time when President's limo appeared at the Dealey Plaza (the curve from Main onto Houston street) occurred at 12:29:21 and this was the time when a number of people at Dealey Plaza started to cheer the incoming motorcade.  This could be the moment which led Lee Oswald to a decision to go out and find out what was this excitement about.

The next significant moment is the time when President's limo passed the Depository entrance and this period is depicted in Hughes film. I spent a lot of time exploring Hughes film and could not find any evidence of Prayer Man's presence during the moment when the motorcade was in front of the doorway. The time when the limo was passing the the entrance of the building was 12:29:46. Finally, Prayer Man is seen in Wiegman film and Wiegman's camera captured the doorway about at the time of the last shot (12:29:59). Thus, Lee Oswald could enter the doorway only during a specific time interval falling between the instant when the limo was in front of the Depository entrance (depicted in Hughes) and the instant corresponding to the final shot (equivalent to the onset of Wiegman's filming of the Depository building). This interval would be 13 seconds, however, since Prayer Man was not at the western wall in Hughes and it takes about 2 seconds to move from the glass door to the western wall, the 13 second interval needs to be cut by 4 seconds (2 seconds after the Hughes event and 2 seconds before the Wiegman event). This results in a 9-second interval during which Prayer Man could occur in the doorway if he was coming from the inside of the building. The 9-second interval in real time would start at 12:29:48 and end at 12:29:57. 

 

Counting backward, basically subtracting the walking time of 24 seconds, it is possible to estimate the time interval during which Lee Oswald could leave Domino room to reach the doorway not earlier than 2 seconds after Hughes and not later than 2 seconds prior to Wiegman's filming the Depository.  This gives the time interval 12:29:24 - 12:29:33.  The earliest moment (12:29:24) occurs 3 seconds after the limo turned from Main onto Houston street and such interval appears sufficient for someone to make a decision. However, Lee had even more time to make up his mind about leaving Domino room  and going out - any instant falling within the 9 seconds after 12:29:24 would bring him to the doorway just on time to be filmed by Wiegman and not to be filmed by Hughes.

[Mathias Baumann]

10 hours ago, Andrej Stancak said:

Returning to the possibility of increasing the resolution of existing Prayer Man images, a more generic approach appears to be deep learning for single-image super-resolution. This has been implemented in Matlab environment: 

https://www.mathworks.com/help/deeplearning/ug/single-image-super-resolution-using-deep-learning.html?s_tid=answers_rc2-2_p5_MLT

It is a convolution neural network which is first trained on many thousand images how to reconstruct a high-resolution image from a low-resolution version of the same image. Only the luminance channel is used in this method which is good because Darnell's stills are black-and-white images.  The network, after being trained on still images of people, animals and various objects, is able to apply the learned network layers to a sample low-resolution image and increase the high-frequency features (e.g, details of the face) of that image. Comparison of the standard bicubic interpolation and the deep learning method showed an increase in similarity with the original high-resolution image from about 0.25 to 0.87 (1.0 being similarity with its own copy).

This is a major project requiring lengthy computations, some reading and iterative improvements. For instance, using a database of containing only human faces could train the neural network better for the purpose of detecting features in Prayer Man's face than variety of objects (animals, landscapes...). I hope to be able to embark on this in the future, however, this may not be a very near future because of a number of parallel projects and commitments.

 

Can this be done with a regular desktop PC or do you need some special hardware?

[Sandy Larsen]

18 hours ago, Andrej Stancak said:

Returning to the possibility of increasing the resolution of existing Prayer Man images, a more generic approach appears to be deep learning for single-image super-resolution. This has been implemented in Matlab environment: 

https://www.mathworks.com/help/deeplearning/ug/single-image-super-resolution-using-deep-learning.html?s_tid=answers_rc2-2_p5_MLT

It is a convolution neural network which is first trained on many thousand images how to reconstruct a high-resolution image from a low-resolution version of the same image.

 

I commented earlier that generic photoshop-type filters cannot be used to properly deblur images... that is, in a way that reveals true image details. Well, my skimming through the above linked article reminded me of something that MAY invalidate what I said. I need to give a bit of background before I explain.

When an image is blurred, electrical engineers and mathematicians model the process by saying that the original sharp image has been blurred by convolving the image with the point spreading function (PSF) of whatever caused the blurring. Similar to a + or - symbol used for addition or subtraction, we use the asterisk symbol for convolution. Thus

SharpImage  *  PSF  =  BlurredImage

If the sharp image were just a point of light, the blurred image would display the PSF (point spreading function).

If we knew what the PSF was, we could do an inverse operation -- referred to as a deconvolvolution -- on the blurred image to restore the sharp image. Just like division can be used to undo a multiplication.

The problem is that we don't know what the blurring PSF of the Darnell image is. And this is why I said a generic filter cannot be used to deblur the image.

Well, there is a concept called "blind deconvolution" where you use an adaptive algorithm to recover a close approximation of the PSF, with which the sharp image can be recovered. Upon reading the Andrej's linked article, it occurred to me that there might be some good blind deblurring algorithms that have been developed since the time I was  involved in this work.

So I googled the words "blind image deblurring filter" and found a lot of useful information. Happily there is a MATLAB function, called deconvblind, for performing adaptive blind deblurring. It is a trivial thing to try for someone with MATLAB and the Image Processing Toolbox (U.S. $200 for U.S. educators; free for 30 days,  I think). This link shows how it can be used and gives a real example:

Deblurring Images Using the Blind Deconvolution Algorithm

Unfortunately it doesn't give the name of the algorithm used for the blind deconvolution (maybe it does somewhere in it documentation), so I don't know how it works. .

This page gives more general MathWorks information regarding image deblurring:

MathWorks - Image Deblurring

Having said ALL that (sorry!), back to my first paragraph... it is conceivable that Photoshop likewise DOES have a blind deblurring function that really does reveal image details. I don't have Photoshop so I can't check.

 

Edited August 30, 2020 by Sandy Larsen

[Sandy Larsen]

Just now, Sandy Larsen said:

Happily there is a MATLAB function, called deconvblind, for performing adaptive blind deblurring. It is a trivial thing to try for someone with MATLAB and the Image Processing Toolbox (U.S. $200 for U.S. educators; free for 30 days,  I think). This link shows how it can be used and gives a real example:

Deblurring Images Using the Blind Deconvolution Algorithm

 

I hope that Andrej, or some other smart person like Andrej, will give that MATLAB function a try. I'm not in a position to do so myself.

 

[Sandy Larsen]

4 minutes ago, Sandy Larsen said:

Having said ALL that (sorry!), back to my first paragraph... it is conceivable that Photoshop likewise DOES have a blind deblurring function that really does reveal image details. I don't have Photoshop so I can't check.

 

I just found out that Photoshop DOES have such a filter, designed to remove blurring caused by camera shake. Whoever has Photoshop can try applying that filter to Prayer Man. Click on Select Filter > Sharpen > Shake Reduction. I would try various image sizes (that is to say, resized by cropping off surrounding areas) to see if that makes a difference.