Jump to content
Sign in to follow this  
SweatyYeti

Patty's Arms And Hands

Recommended Posts

SweatyYeti

Here is something else, Neander. Using the 'foot ruler' in Frame 61...(a better view of the foot than in F72)...this is what I get for Patty's arm length...

 

 

F61-FootRuler-ArmLengthMeasure1_zpse5e08

 

 

It measures a few inches longer than your 'average Bob's' arm.

 

One interesting consequence of that, is how would Bob have been able to make the fingers bend so close to the end of the fingers....(clearly beyond the end of his fingers)???...

 

th_PattyFingersComp2_zps5a57a98c.jpg

Edited by SweatyYeti

Share this post


Link to post
Share on other sites
MikeZimmer

Fitting Vision Realm Entertainment Skeleton into Patty Image

I did some work trying to see how the skeleton image from Vision Realm Entertainment Incorporated, produced for the Sasquatch, Legend Meets Science video, fits in with frame 352 from the Patterson-Gimlin film. I used the false colour version, taken from work by MK Davis. I previously had taken measurements, and computed various anthropomorphic indicies. I heard from several people that there were flaws in the methods and the measurements. I decided to see it I could harmonize the Vision Realm leg and arm proportions with those seen in frame 352.

 

I believe that this is permissible under the laws of Canada, Fair Dealing, and the US, Fair Use.

 

Frame 352 is nice to work with because it presents a view that appears to have minimal foreshortening. However, it has some deficiencies as well. The first is that both feet are hidden, so the relative length of the lower leg is not available for scrutiny. The second is that the right arm is straight, so it is hard, if not impossible, to determine from this frame alone where the elbow joint is situated. In order to remedy these problems, it will be necessary to use multiple frames for analysis, and not simple methods based on the dimensions of one frame. In particular, the hip and shoulder joints need to be determined by analysis of multiple frames to make note of the rotation about the joints.

 

The proportions of the skeleton do not seem particularly congruent with the creature in frame 352. If the leg is right, then the arm is wrong. If I rescale to make the arm right, at least in length, then the humerus to radius ratio will still be wrong according to other work, and the hip or knee may well be misplaced. There are several possibilities for error that I can think of, although I would probably need independent counsel to help identify the problems.

  1. The work by Vision Realm Entertainment is flawed
  2. The image that I used, reproduced from a book, contains some sort of distortion which is throwing off my measurements
  3. There is some distortion in the aspect ratio of the frame 352 image that I am using which is throwing things off. I do not know exactly how to determine if I have an image with the correct aspect ratio.
  4. My measurements of the bones were inadequate
  5. My initial placement of the knee and hip joint on the right leg in frame 352 were not right. I am reasonably confident in the knee placement, but the hip needs to be independently vetted.
  6. Since the scaling was done with respect to the initial placement of the hip and knee joints, and measurement of the femur on the Patty image, everything else will change if this dimension is changed. Things may still not fit, but the misfit would look different.
  7. My calculations were wrong, my scaling was incorrect
  8. My measurements to place lines on the Patty image were not well done

Here are the images, and supporting work in brief.

  1. At the top is a section on measuring bones from ‘An Introduction to Human Evolutionary Anatomy’, Leslie Aiello, Christopher Dean. Note that I did not precisely follow their method, because I had forgotten about it when first making my measurements.
  2. The second part is a reduced scale image of the Vision Realm Entertainment skeletal image.
  3. The third part is frame 352, with lines showing length, crosses showing end-points, and circles showing a 30 pixel range.
  4. The fourth part is a spreadsheet, with the scaled dimensions shown.

vision-realm-entertainment-image-fitting

Share this post


Link to post
Share on other sites
Guest

It will take some time to get up to speed on the many pages but I like what I'm reading so far. It has already been established the arm is longer than Bob Heironimus's arm. He's the most probable perpetrator and arm extensions don't cut it with the finger movement seen in the film, when the shoe doesn't fit it is very telling. 

Share this post


Link to post
Share on other sites
Guest

Fitting Vision Realm Entertainment Skeleton into Patty Image

I did some work trying to see how the skeleton image from Vision Realm Entertainment Incorporated, produced for the Sasquatch, Legend Meets Science video, fits in with frame 352 from the Patterson-Gimlin film. I used the false colour version, taken from work by MK Davis. I previously had taken measurements, and computed various anthropomorphic indicies. I heard from several people that there were flaws in the methods and the measurements. I decided to see it I could harmonize the Vision Realm leg and arm proportions with those seen in frame 352.

 

I believe that this is permissible under the laws of Canada, Fair Dealing, and the US, Fair Use.

 

Frame 352 is nice to work with because it presents a view that appears to have minimal foreshortening. However, it has some deficiencies as well. The first is that both feet are hidden, so the relative length of the lower leg is not available for scrutiny. The second is that the right arm is straight, so it is hard, if not impossible, to determine from this frame alone where the elbow joint is situated. In order to remedy these problems, it will be necessary to use multiple frames for analysis, and not simple methods based on the dimensions of one frame. In particular, the hip and shoulder joints need to be determined by analysis of multiple frames to make note of the rotation about the joints.

 

The proportions of the skeleton do not seem particularly congruent with the creature in frame 352. If the leg is right, then the arm is wrong. If I rescale to make the arm right, at least in length, then the humerus to radius ratio will still be wrong according to other work, and the hip or knee may well be misplaced. There are several possibilities for error that I can think of, although I would probably need independent counsel to help identify the problems.

  1. The work by Vision Realm Entertainment is flawed
  2. The image that I used, reproduced from a book, contains some sort of distortion which is throwing off my measurements
  3. There is some distortion in the aspect ratio of the frame 352 image that I am using which is throwing things off. I do not know exactly how to determine if I have an image with the correct aspect ratio.
  4. My measurements of the bones were inadequate
  5. My initial placement of the knee and hip joint on the right leg in frame 352 were not right. I am reasonably confident in the knee placement, but the hip needs to be independently vetted.
  6. Since the scaling was done with respect to the initial placement of the hip and knee joints, and measurement of the femur on the Patty image, everything else will change if this dimension is changed. Things may still not fit, but the misfit would look different.
  7. My calculations were wrong, my scaling was incorrect
  8. My measurements to place lines on the Patty image were not well done

Here are the images, and supporting work in brief.

  1. At the top is a section on measuring bones from ‘An Introduction to Human Evolutionary Anatomy’, Leslie Aiello, Christopher Dean. Note that I did not precisely follow their method, because I had forgotten about it when first making my measurements.
  2. The second part is a reduced scale image of the Vision Realm Entertainment skeletal image.
  3. The third part is frame 352, with lines showing length, crosses showing end-points, and circles showing a 30 pixel range.
  4. The fourth part is a spreadsheet, with the scaled dimensions shown.

vision-realm-entertainment-image-fitting

 

 

Hi Mike,

 

Placing the joints accurately requires a human surrogate to establish a relative standard. You can positively identify where your own joints are on a surrogate overlaid with known dimensions. You don't have to be spot on the joint as long as you are consistent. You can rescale your joint estimates when you are ready to compare to human indices. 

 

Since this is an elbow/knee issue, it doesn't affect the total length of the leg/arm significantly. It only relates to the humeral/radial, femoral/tibial indices which requires accurate placement of the elbows and knees. Why not start with the arms/legs to height ratios first and locate the elbows/knees next? We need to get past "arm extensions" first before we can locate the elbow.

 

Funny how the PGF skeptics get quiet when it comes to measuring Patty's body dimensions on film. They just don't like the opinions of the photogrammetrists that have examined the film thus far. They prefer to focus on some nefarious thing Roger did.

 

GF

Edited by Gigantofootecus

Share this post


Link to post
Share on other sites
MikeZimmer

Here is something else, Neander. Using the 'foot ruler' in Frame 61...(a better view of the foot than in F72)...this is what I get for Patty's arm length...

 

 

 

 

 

It measures a few inches longer than your 'average Bob's' arm.

 

One interesting consequence of that, is how would Bob have been able to make the fingers bend so close to the end of the fingers....(clearly beyond the end of his fingers)???...

 

 

 

 

Unless I am mistaken, it is foreshortened as well, as shown in a previous calculation by Gigantofootecus, so it is even longer.

 

Mike

Share this post


Link to post
Share on other sites
Guest

So many people think the video was a fraud, if they only took the time to really look at all the facts.

Share this post


Link to post
Share on other sites
MikeZimmer

Hi Mike,

 

Placing the joints accurately requires a human surrogate to establish a relative standard. You can positively identify where your own joints are on a surrogate overlaid with known dimensions. You don't have to be spot on the joint as long as you are consistent. You can rescale your joint estimates when you are ready to compare to human indices. 

 

Since this is an elbow/knee issue, it doesn't affect the total length of the leg/arm significantly. It only relates to the humeral/radial, femoral/tibial indices which requires accurate placement of the elbows and knees. Why not start with the arms/legs to height ratios first and locate the elbows/knees next? We need to get past "arm extensions" first before we can locate the elbow.

 

Funny how the PGF skeptics get quiet when it comes to measuring Patty's body dimensions on film. They just don't like the opinions of the photogrammetrists that have examined the film thus far. They prefer to focus on some nefarious thing Roger did.

 

GF

 

 

Hi Gigantofootecus,

 

I bought an artist's mannequin a couple of weeks ago, and also created a skeletal model as suggested by you, with the key bones and points of articulation. I could not find the original suggestion from you, so I made up my own interpretation. For arguments sake, I used the Vision Realm Entertainment skeletal proportions, knowing that they probably will not hold up. I am still trying to figure out where to go with that one.

 

I also have been using myself as a human mannequin. I suppose the next step is to make some videos of human movement for calibration purposes. I can locate the joint centres with palpation, and have found videos on how to do this.

 

I put that stuff on hold, as I started intensively developing methods for locating joint centres using geometrical methods. I wrote up a paper on the techniques I was working on at the end of February. See article here: Measuring Human Long Bones With Known Bounds For Uncertainty

 

I have since then developed better methods for determining rotation centres from multiple frames. I have been testing them on some video frames of human runners. I still don't have good quality Patty images. I will see what I can do about that when my latest methods are properly tested and documented.

 

The latest method I have come up with seems to be very forgiving of error, revealing the proper centre on test examples. Another one I developed yesterday was a bust, but the one the day before, requiring three more-or-less-sequential frames was pretty good. The newest only needs two frames. More can be used for validation, and for alternative estimates.

 

The method essentially constructs a line along the outside margin of the anterior thigh, and a similar line along the posterior thigh. These are extended so they meet at a vertex. This vertex angle is subdivided into four identical divisions, quadrisected. This is done for two frames. Using a nice little graphical algebra and geometry tool called GeoGebra, I then combine the frames as lines in Cartesian space. I construct angle bisectors where the anterior lines meet, from the equivalent angle division, and where the posterior lines meet. Where these new bisectors meet is the estimate of the centre of rotation. This method is rigorous for pure geometry, but also seems to be pretty tolerant of errors in line construction. The frames do have to be registered based on some identifiable surrogate for the centre of rotation (you don't actually know where that is). The errors can be due to:

  1. registration (scaling,translation, and rotation)
  2. construction of the lines along the thigh margins, with somewhat indistinct margins
  3. perspective
  4. depth of the joint centre - the hip and shoulder bones are of course within the body, and with differing perspectives, even with perfect registration based on surface features, there will be some error because of the angle of the shot. There may be a way to compensate for that, but first let me make sure the basic method works.

I need to validate the methods more thoroughly, and document so that they can be followed. I want to explain why they work, and how well they work, giving an analysis of sensitivity to erroneous data values. That will entail a lot more work.

 

After that, I hope to obtain some very good quality images, centering around the look-back frame. There are some very nice surface markings around the hip bone in the Cibachrome, and also in the MK Davis false colour work. I am hoping that these markings show up in associated frames. The frames I have right now are just not good enough I feel. If it is possible to get several frames of near Cibachrome quality in that section of the film. I will proceed with trying the methods out.

 

In my work with videos of runners I have registered at the mid-point of the belt-line. This is only approximate, and I would like to find a better surrogate for the centre of hip rotation.  With Patty, I am hoping as I said to use surface markings. I have seen cases where the error was fairly great where translation was inaccurate, and another case where I somehow ended up with an image incorrectly scaled.

 

In one runner video, I have 15 frames encompassing a full (two step) stride. In my look at the Patty movie, I don't see as many frames for a full stride. This is a mystery to be resolved later. The frames per second should not differ that much video to film, and if anything, there should be fewer frames/stride for the runner.

 

The main runner video was shot from the side of the road, so the foreshortening is pretty bad on some frames. The method will work best where the perspective is roughly constant. With images shot at 100 feet, I believe that the problem is smaller. I have only used the method when the shot is lateral, mostly perpendicular with the plane of the lens. I would imagine the results will be less correct at other angles.

 

I don't know if I have reinvented a wheel or two here or not. Usually, I am not the first with my ideas. However, I suspect that not a whole lot of people have tried to do this with images.

 

I will start to consider 3D geometry with the mannequin and other things later.

 

Mike

Edited by MikeZimmer

Share this post


Link to post
Share on other sites
Guest

So many people think the video was a fraud, if they only took the time to really look at all the facts.

 

These people either just don't want to know or they refuse to accept that the obvious facts right in front of them don't exist. It helps them sleep better at night.

Share this post


Link to post
Share on other sites
MikeZimmer
...

 

The method essentially constructs a line along the outside margin of the anterior thigh, and a similar line along the posterior thigh. These are extended so they meet at a vertex. This vertex angle is subdivided into four identical divisions, quadrisected....

 

 

I got rid of the quadrisector idea (above). In my three frame method, it adds accuracy, in this two frame method, it does not, since the estimated centre points turn out to be the same. Not true of the three frame construction. Still getting my head around why.

 

I applied it to Bill Munns' model, for a single frame only, and it still gave good results. I have posted a snapshot below, but if the moderators believe it inappropriate, please remove it. It is nothing that you would not see on any beach of course. I hope that Bill does not mind me using a bit of his work, in the name of research.

 

I will be writing up a description and explanation at some future date.

 

 

 

bill-munnss-model-single-frame-rotationa

Edited by MikeZimmer

Share this post


Link to post
Share on other sites
MikeZimmer

I got rid of the quadrisector idea (above). In my three frame method, it adds accuracy, in this two frame method, it does not, since the estimated centre points turn out to be the same. Not true of the three frame construction. Still getting my head around why.

 

I applied it to Bill Munns' model, for a single frame only, and it still gave good results. I have posted a snapshot below, but if the moderators believe it inappropriate, please remove it. It is nothing that you would not see on any beach of course. I hope that Bill does not mind me using a bit of his work, in the name of research.

 

I will be writing up a description and explanation at some future date.

 

 

 

bill-munnss-model-single-frame-rotationa

 

 

Here is the single frame tangent bisector method applied to frame 352 in false colour, produced by M.K. Davis. I expect that the true centre is just slightly lower and anterior to this estimate, because the left leg and right leg joint centres are probably not perfectly aligned. This is because the hips are almost certainly slightly turned. If the joint centres were properly registered, the results would be darn close to perfect.

 

patty-frame-352-rotational-tangent-metho

 

 

Compare this to previous estimates by myself, using guesswork guided by an anatomical model, and by Gigantofootecus, using his methods.

The advantages of using this method are:

1 - it is probably going to be shown to be quite accurate, based on preliminary validation results

2 - it is defensible, in that it can be shown why it works, and the factors affecting accuracy and uncertainty can be qualified, and quite possibly quantified.

 

mkdavis-gigantoofootecus-comparison.png
MK Davis Patty Frame 352 Hip Centre Location
Edited by MikeZimmer

Share this post


Link to post
Share on other sites
Guest

Hi Mike, above and beyond the call of duty!

Dusted off your old geometry textbooks I see. Just make sure all your geometry lines up with your biology. Great to hear that you got the mannequin and you are willing to measure your own images! I know this is a tall order, but you must film the mannequin with a ~25mm lens at ~100 feet from the camera AGAINST A GREEN SCREEN! Pain in arse, I know, but you need to eliminate the effects of a "close to the camera image", which all the CGI images seem guilty of, and overlay your images over the PGF. You need a green screen mask for that.

Also, put white balls on all the joints on your mannequin, measure the distances between them and use these dimensions for your initial model. Have 1 camera zoomed in on your mannequin, which is positioned to match a frame from the PGF, which you display on a monitor beside the monitor of your mannequin image. When you get these images "squared" you can create an overlaid image of the mannequin over the PGF frame in a blinking GIF. This will help you determine whether you have aligned the joints correctly. If not then adjust the model and repeat.

In the end you want the same results as Vision Realm. A series of frames of your mannequin mimicking Patty with the same dimensions. The big question is, are your mannequin's dimensions adjustable?

Cheers

GF

Edited by Gigantofootecus

Share this post


Link to post
Share on other sites
MikeZimmer

Hi Gigantofootecus,

 

I don't know the lens on my camera, so will have to investigate that. Getting a 100 foot distance means going outdoors. Since the mannequin is only 12 inches, is it not going to be ridiculously small in the image? I have a high def camera, but will it be able to resolve sufficient detail?

 

What do you use as a green screen? Will green Bristol Board do?

 

I wanted to put the mannequin on a turntable calibrated in degrees so I could investigate different known angles.

 

The mannequin is cheap, with no ability to change the dimensions. I toyed with making one from plastic pipe, but that seemed to be overly difficult, with some technical solutions required that I have not thought through. Telescoping pipe would work, it I could find the right diameters - a snug slip fit. Also, I am not sure on how to hinge things - ball and socket especially.

 

Mike

Share this post


Link to post
Share on other sites
Guest

Thinking about it Mike, with tilting and twisting the mannequin and with all the different camera positions there's so many possibilities of different views. 

Share this post


Link to post
Share on other sites
MikeZimmer

Thinking about it Mike, with tilting and twisting the mannequin and with all the different camera positions there's so many possibilities of different views. 

 

 

Agreed. I am hoping to get, at the very least, a better feel for how the apparent geometry changes with differing perspectives, and maybe get some clearer insight into how to compensate for the foreshortening. This is all new territory for me, and takes me from 3D geometry into 3D. I did not learn much of that.  The issue about the 100 foot distance is that the effect of foreshortening becomes quite small at that distance.

you can also quantify it with some very simple ratios regarding triangles, taken from trigonometry. My older thread on measurement, http://bigfootforums.com/index.php/topic/50200-measurement-gotcha%E2%80%99s-%E2%80%93-a-neophyte%E2%80%99s-guide/

 

 

Regards

  • Upvote 1

Share this post


Link to post
Share on other sites
Sign in to follow this  

×
×
  • Create New...