Jump to content
SweatyYeti

Patty's Arms And Hands

Recommended Posts

SweatyYeti

^

 

It always looked that way to me, too, Neander...until just a couple of days ago. :)

 

In working with this image, I finally noticed the bottom of Patty's right foot...

 

SO-MiddleSequence-F5G_zpsf830e96b.jpg

 

 

One of the strong points of this particular copy of the Film is it's color quality....it looks near-perfect. The ground appears light-to-medium gray...(as it should)....and the bottom of Patty's foot appears very similar...on the grayish side.

Edited by SweatyYeti

Share this post


Link to post
Share on other sites
Backdoc

^^^

Even though apes and people have lighter colored palms and feet bottoms, it is reasonable to think the Patty figure has Bluff Creek soil on the bottom of those feet.  Too much has been made of the 'white' bottoms of the feet when they clearly are not 'white'     Also, we would have to assume such and excellent suit has such a glaring weakness like white feet.

 

Backdoc

Share this post


Link to post
Share on other sites
MrSkwatch

Is it possible for the original film to be clear enough to show tracks being left by Patty as she walks?

Share this post


Link to post
Share on other sites
SweatyYeti

^

 

I don't think the tracks can be seen being made at any point in the Film, Skwatch.

 

During the steady 'look back' segment Roger was filming while down on his knees...so the camera is too low to get a view of the footprints. The image I posted above was taken shortly before Roger fell to his knees...and the camera is high enough to at least catch Patty's shadow on the ground.   :)

Edited by SweatyYeti

Share this post


Link to post
Share on other sites
MikeZimmer

^

Sweaty,

 

What do you think caused the two vertical parallel lines in the frame?

Share this post


Link to post
Share on other sites
SweatyYeti

MikeZimmer wrote:

 

Sweaty,

 

What do you think caused the two vertical parallel lines in the frame?

 

 

Those are just scratches from a projector, Mike. That's all... :)

Edited by SweatyYeti

Share this post


Link to post
Share on other sites
MikeZimmer

Hi Mike,

 

You must derive the articulation points. The easiest way to do this is to go old school and use an adjustable mannequin and develop a model by matching Patty's orientation and distances from the camera for frames 288 to 480. Adjust the dimensions as required to fit the model. This frame sequence represents 16.8 steps of the clearest, most consistent film footage and Patty is essentially a walking ruler @ approx. 12 frames per step.

...

 

Happy New Year!

GF

 

 

Dang, I may be starting to understand what you are saying now Giganto, after doing a lot of work coming to grips with some of the fundamental issues.

 

It is interesting that the eye can see things in a moving image that become very hard to pin done when looking at things on a frame by frame basis. The edges of objects in static frames are a lot harder to see than when they are seen in the context of the video. Our perceptual system does some pretty amazing processing.

Share this post


Link to post
Share on other sites
MikeZimmer

Some geometric techniques that I have been experimenting with. I have had some difficulty trying to apply them to Patty, but I almost certainly don't have the best images.

 

Graphically Determining the Elbow Joint Placement From External Image Contours

 

 

This study uses a couple of techniques for creating landmarks to determine the centre of rotation for an image of a somewhat flexed elbow. I used X-ray images in the work here, so that the accuracy of the geometric constructions can be assessed. It works on lateral views of an elbow. It may work with lessened accuracy on oblique views, but this is still a matter for investigation. It is not clear yet if the results are affected by foreshortening. My current intuition is that it will be a problem for large amounts of foreshortening, hence the reservations about oblique views.

 

The method demands that the elbow be bent. The technique is impossible with a straight arm, and surely increases in accuracy with more sharply bent arms, at least up to a point. This is a matter for investigation.

 

It is not clear if this gives greater accuracy than joint positioning by other methods, such as “by guess and by gollyâ€

 

Two Methods:

 

Two geometric methods are shown:

  1. Using Inner and Outer Elbow Landmarking
  2. Using Outer Elbow Landmarking and Angle Bisectors

Both demand that a reasonable outline of the upper and lower arms near the elbow joint can be seen.

  1. The first method demands that both inner and outer aspects of the arm are seen.
  2. The second only asks that the outer aspect is seen.

The positioning of the joint is inexact, but the construction lines should come close to crossing the centre of rotation with respect to the length of the arm.

 

The positioning with respect to the width or the arm can not be determined in general, since arms will differ greatly in the amount of soft tissue surrounding the bones. An emaciated victim of a wasting disease will vary greatly from a very thick bodied person, either muscular or fat. In general, on an average arm, the actual centre will be closer to the outside of the elbow than the inside, by a considerable margin. If this technique is found to be workable for images lacking in some clarity, it might be possible to determine some function based on arm length and circumference to see where the centre lies with respect to the outer and inner aspects.

 

Note that the centre of rotation does not define the point of the end of humerus, radius or ulna, since they are a bit off from that point. This is shown in the third image.

 

Using Inner and Outer Elbow Landmarking

 

elbow-heavy-curvature-multiple-lines.jpg

 

Using Outer Elbow Landmarking and Angle Bisectors

 

bisector-method-trial-1.png?w=700&h=700

 

Placement of the Long Bone Endpoints with Landmarking

 

determining-the-elbow-articulation-from-

 

 

 

Source at https://viewfromsitchamalth.wordpress.com/2015/02/02/graphically-determining-the-elbow-joint-placement-from-external-image-contours/

Edited by MikeZimmer

Share this post


Link to post
Share on other sites
MikeZimmer

I tried working an example for my angle bisector method for locating the elbow joint (see above). This post is taken from the source document at https://viewfromsitchamalth.wordpress.com/2015/02/03/bisector-method-on-patty-animated-gif/. I hope that my usage of Sweaty's GIF is kosher. I have been told that I might be able to get images of frames from elsewhere, but I have not asked elsewhere yet.

 

I am not sure that my construction line on the forearm is not on a shadow instead of an arm contour. If I got it wrong, the method I used disagrees with Sweaty's to some as yet undetermined extent. Unfortunately, blowing up the image does not make anything clearer. The debris in the foreground does not help.

 

Bisector Method on Patty Animated GIF

 

Look for the last image in the animated GIF to see my placement of the elbow. I managed to use the bisector method to locate the elbow. The red lines are my construction of landmarks, and the yellow dot is SweatyYeti’s estimate.

 

pattyarmbendproportionscompag2mod.gif?w=

 

I don’t understand how to use animated GIFs  with the GIMP software, so I could not work successfully with other layers of the GIF. There is another frame that looks as though I could use the double “knee†method

.

Since the images are poor quality, there is still some “guess and by golly†in deciding upon the contour or the arm. However, despite that, I arrived at almost the same elbow joint placement as did SweatyYeti. I may have erred a lot, but somehow I ended up with a very similar result.

 

I am hoping that this is not the best quality image that is out there. I took something that SweatyYeti had previously created.

 

The angle bisector construction is a pain, since I don’t have a tool to calculate it for me. I used a spreadsheet, and it took some thought to figure out how to calculate it, since GIMP gives the angle for two points only with respect to a particular quadrant. Add this, subtract that, take the 90s complement of something else. Made my brain hurt. I could have done it more readily with paper and drawing compass.

 

angle-bisector-calculations.png?w=300&h=


Some geometric techniques that I have been experimenting with. I have had some difficulty trying to apply them to Patty, but I almost certainly don't have the best images.

 

...

 

 

Placement of the Long Bone Endpoints with Landmarking

 

 

 

Source at https://viewfromsitchamalth.wordpress.com/2015/02/02/graphically-determining-the-elbow-joint-placement-from-external-image-contours/

 

Apparently I need to explain things in more detail. The red, green and blue lines are independent estimates of where the centre of rotation of the elbow joint lies. Note that each set is based on different takes on the contours of the flesh of the arm. There is quite a good agreement, and all cross very near the centre of rotation around the humerus. The last image also shows where the humerus, ulna, and radius all terminate, by using short black lines at the end-points.

Edited by MikeZimmer

Share this post


Link to post
Share on other sites
MikeZimmer
....

Bisector Method on Patty Animated GIF

 

Look for the last image in the animated GIF to see my placement of the elbow. I managed to use the bisector method to locate the elbow. The red lines are my construction of landmarks, and the yellow dot is SweatyYeti’s estimate.

 

...

 

 

Looking at this method, I am starting to realize its deficiencies, as well as where it might be necessary.

 

It occurred to me that I could apply the bisector method for estimating the elbow (or knee) joint centre of rotation from the inner angle of the joint or the outer angle. I could not initially visualize the answer. Do the inner and outer angle bisectors intersect, or are they congruent? I expected the latter in the case of perfectly measurable angles, such as on a swinging door, or a drop -leaf table with hidden hinges, but expected that the inaccuracy from constructing landmarks on a real creature would mean that they are not congruent. I thought some more, did some image work, and came to a firmer understanding.

 

If the arm segments were of uniform width along their length, then the bisectors, taken either on the interior angle of the elbow or the exterior angle of the elbow, would be fully coincident. However, since the arms are not uniform along their length, the results from interior bisector and exterior bisector will not coincide. In addition, there is more error with this method than with the interior and exterior vertex method.

 

Where you would use it would be when you actually can only clearly define one vertex, either interior or exterior. It might be possible to determine to some level of precision how far off the true centre the bisector is likely to be, but probably it would rely on some assumptions which themselves might not be solid.

 

It is all a matter of angles and geometry.

Edited by MikeZimmer

Share this post


Link to post
Share on other sites
MikeZimmer
...

 

It is all a matter of angles and geometry.

 

Analysis of Methods for Locating the Elbow from an Image

 

When looking at the right arm of the creature in the Patterson-Gimlin film, it is not totally obvious where the elbow joint is situated. The images are not incredibly clear, so details are not necessarily visible. Sometimes they are partially obscured, and frequently they are taken from an oblique perspective.

 

Various methods have been used by others to estimate the position, but all attempts have been called into question at various times. The issue is one of accurate placement, and in addition, making the method of placement clear so that it can be rationally debated. There will always be uncertainty, but that uncertainty should be quantified to the best of our ability.

 

This is a test of three systematic methods of estimating the elbow articulation based on geometric techniques, on single image frames. With reliable methods for the shoulder, elbow and wrist, and for the hip, knee, and ankle, it will be possible to get measurements of humerus, radius, femur and tibia, and compute some standard ratios and analyze statistics for these ratios. Other have done some of this, but it is not clear in all cases what methods were employed, and if the methods were sound. There does not yet seem to be an accepted and complete set of measurements.

 

The following methods were used to determine where the bones of the arms articulate at the elbow joint. The methods were based on geometrical constructions and landmarks. Various X-ray images of the flexed elbow, taken from a lateral perspective, were analyzed using the image processing software GIMP. By using X-ray images, it is possible to determine how effectively the technique works.The methods explored to determine approximate position of the elbow joint were:

 

1 – using the inner and outer elbow vertex construction to determine position

2 – using the inner and outer elbow vertex with angle bisectors

3 – using long axis bisection or upper and lower arm

 

None of these methods gave precisely accurate results, but they are listed above in order of decreasing accuracy.

None of the methods could give the precise centre of rotation of the joint, although method 1 gave a reasonable method of determining the position of the centre with respect to the long axis or the arm. None of the methods could give the position of the centre across the width of the arm, and in fact, the fleshiness of the arm would have to be taken into account to determine that.

 

A fairly good approximation of the true centre of rotation was made by eye, based on the circular contour of the end of the humerus bone. It was found that the end of the humerus and the end of the radius bones are not at the centre of rotation, and some adjustment would have to be made to get a more accurate positioning.

 

The samples were all of fairly well-flexed elbows, since lateral views of less bent elbows were not found. In general, X-rays of the elbow joint are taken with the elbow bent, so other images may be hard to find.

 

 

Left lateral view of elbow, with construction lines

 

 

three-methods-elbow.png?w=300

 

Legend:

 

Red lines: using the inner and outer elbow vertex construction to determine position

Blue lines: alternate for using the inner and outer elbow vertex construction to determine position

Green Lines: using the inner and outer elbow vertex with angle bisectors

Black lines: using long axis bisection or upper and lower arm

Line A: end of the radius bone

Line B: end of the humerus bone

Point C: close approximation of the centre of rotation

 

Note that no method fixes the centre of rotation exactly

 

Computation of adjustment factors for the end-points of the bones, and error bounds is a work in progress. Since I do not have an X-ray image of the whole arm, from the lateral perspective, I will have to use indirect methods to find adjustment factors. A number of images will have to be subject to analysis to get a clearer idea of the variability inherent in these geometrical techniques.

 

My impression is that they are reasonably accurate, but further study is required.

Edited by MikeZimmer

Share this post


Link to post
Share on other sites
MikeZimmer

Error Reduction

 

 

I have been doing some image analysis work on the false colour images, produced by MK Davis from materials he obtained under copyright from Patricia Patterson: https://www.youtube.com/watch?v=7ve7Fww3ilQ

 

These include detailed panning over the right thigh, knee and hip, torso and head – but not the whole image.

 

Here is my application of geometric construction and measurement to a Patty image of the hip and upper leg from the false colour work by MK Davis. On the left are my geometrical lines, on the right are those done by our pseudonymous Gigantofootecus from the Bigfoot Forums. My actual measurement between the centres of the red circles is given in the spreadsheet below, last line of numbers. The images, left and right, are not scaled to be commensurate. I have not figured out how to do this yet.

 

I have included the MK Davis partial image in the belief that this use is legitimized under the "Fair Dealing" provision of Canadian Copyright legislation and the "Fair use" provision of American copyright legislation. If there are any objections, I ask the moderators to remove them.

 

 

mkdavis-gigantoofootecus-comparison.png?

This stuff brings up contrast and detail that just can not be seen in the regular colour. It is amazing – have a look if you have not seen it.

What I was able to pull off the Internet from the MK Davis video for the whole frame itself was OK in quality, but clearly MK Davis has much better stuff. Look at the resolution of his pan over parts of Patty and you can see that there is so much more detail – much better to work with. I want that quality of image!

 

How the heck do you get images of that quality? Gigantofootexus suggested in one post that I contact Bill Munns. I will have to read Bill Munns book again. Stuff does not stick if you have no reason to believe it relevant to your efforts.

 

Thinking about the trail of transformations of the image information from creature to the images I have on my desktop is quite an exercise. So many steps, and so much potential for distortion. It is like computers – when you understand them at all levels, hardware, software, layer on layer, it is not surprising that there are bugs, it is more surprising that the whole Rube Goldberg scheme works at all. Ditto for the images on my machine.

 

My own section of the chain: Windows 7 > Firefox > Video Download Helper > DeVeDe > Pinnacle Studio 11 > Desktop. I tried Video Lan for capture but it gave very low resolution, and I did not locate a setting for bumping that up. Convenient? Yes. Good quality? No. I might try Snagit later.

 

Using the frame (#352 from memory), I came up with a very close agreement with Gigantofootecus on the endpoints of the leg. My points were very close to Gigantofootecus’s, but on the knee, I went more posterior than did he, based on my reading of an anatomical diagram. I have not found a true X-ray side view of the knee, the shots seem to be oblique. The shoulder is way worse though. I also put some error bounds (circular) around them – still unjustified in terms of really quantifying uncertainty, but I made them 15% of the total length.

 

Not being a lawyer, I am not sure about the copyright implications. I think that under Canadian law, I am pretty safe under the “Fair Dealing†provision. I think that under U.S. law, others are pretty safe under the “Fair Use†provision. I would not expect any issues to arise, but I like to follow the rules.

I also don’t want to alienate others such as MK Davis, and ultimately the copy right holder.

 

I did some work on circular error, reproduced below:

error-bars.png?w=423&h=330

The vertical lines are 300 pixels long. We are dealing with images in an X-Y coordinate system, so the errors of placement of the end-point are circular. The size of the circle, diameter measured in pixels, is calculated as error bounds. The upper circle represents the error bounds at one end of the measurement, the lower circle represents the error bounds at the other end of the measurement.

 

The spreadsheet image gives some calculations for error bounds – limits and percentages. I have no idea what odds I would assign on any of these. Probably no better than 50/50.

 

error-calculator.png?w=434&h=186

 

I would like to be able to say with confidence that I was at 95% probability, and within +/- 5%, but that may beyond anyone’s abilities (even on a clear image of a person). The landmarks are inexact, humans vary at the skeletal proportion level, and we are dealing with a purported non-human in the case of Patty.

 

Did I mention that the methods used by science on cadavers or on humans are themselves subject to error? Measurement of bones is probably the most reliable, but even there there are different schools of thought on where to place the ruler.

 

Remainder at: https://viewfromsitchamalth.wordpress.com/2015/02/06/false-colour-images-and-error-reduction/

Edited by MikeZimmer

Share this post


Link to post
Share on other sites
MikeZimmer

I just found this "PGF - ' Full Frame ' Images", which I had not looked at since well before I had any interest in images. Time to read it more carefully, since it may answer some of my questions.

 

#1 icon_share.png SweatyYeti

Share this post


Link to post
Share on other sites
SweatyYeti

^

 

Very interesting stuff, Mike. I'll have to take a day, or two...or three....to look it over, before I can give you much feedback on it. :)

Edited by SweatyYeti

Share this post


Link to post
Share on other sites

×
×
  • Create New...