- Nov 4, 2013
- 16,413
- 1,862
- Gender
- Male
- Faith
- Christian
- Marital Status
- Private
Ok so despite my objection to having to do all the explaining and analysis for the links I posted. Because its already there and explains all the objections raised so I figure its up to the skeptic to research this themselves.
But nevertheless I dug out sections where I think this is relevant to the objections. Here Karoyl explains the software used to analyse the vase. Despite the ad hominems about these researchers being amateurs the software developed to handle these unusual vases as precedents in precision metrology had several experts behind it.
Professor Marian Marcis PHD: Photogrametry, image scanning, 3D reconstruction and digitalisation of cultural heritage.
Johannes Bjorn Meyer PHD: Mathmatics, Geometry on negatively curved spaces, Signal Processing, and Medical research Engineering.
Marton Szemenyei PHD: Electrical Engineer, Computer Vision and Deep Learning research, Ai and 3D in Robotic perception.
In fact they are pioneering new software and methods for detecting these unusual and unprecedented precision vases.
From about the 28 minute mark Karoyl Poka who also has a PHD in Electrical Engineering and computer science explains how the software was developed and how it does render 3D models for analysis.
At the 29.20 minute mark it explains the software called Petriescope and how it works in cleaning up images of unnecessary obvious noise.
Segmentation.
Once clean it is exported as a STM Mesh. The realignment is done using Principle Coordinate Analysers (PCA) which analyses all the points in the mesh to find its main orientation of the vase for alignment. From there an automatic segmentation algorythm identilogue is created of all vase sections or points. It only needs two imputs (top and bottom vertices) on the vase and it generates the required vertex groups. This step lets allows testers to isolate intersections while ignoring chips and missing pieces.
Alignment.
Set the vase straight in the grid relative to the global Z axis. The core of the alignment process in Petriescope uses two algorithms. One alogorithm slices the vase in 3D layers as did Maximus.
Each slice is measured for circularity against the best fit circle to the cross section outline of the vase. Then connects the center of all those cross section circles effectively tracing the objects central axis from bottom to top. The estimated axis is then aligned to the global Z axis.
Analysis
Mainly focuses on circularity or how perfectly round a cross section is and concentricity which means how well centered the cross sections are to the global axis Z.
For circularity the software slices align the vase into many horizontal sections almost like a Cat Scan but are a mesh. For each slice it fits a perfect circle to the outline of the cross section. Then it calculates a value called Root Means Square Deviation (RMSD) for that slice. In simple terms RMSD is like an average error. It measures how far each cross section deviates from the perfect circle. Each slice gets its own RMSD value telling us the level of roundness in that slice of the vase.
Then the single metric or (Median) is used for the entire vase as explained earlier. It is the best method as it allows for chips, damage or scanning noise outliers that may skew the findings.
Standard geometric dimensioning and tolerances practice or GD & T often uses a high to low median for roundness. Basically the difference between the maximum and minimum radius in a cross section.
Then placing these circles in a CAD model as the perfect reference points. This produces color coded heat maps that reveal the true surface deviations. This offers a full 3D picture graph view rather than a flat 2D view arrow plot.
Industrial inspection normally compares parts to the original CAD drawings. Since we lack 5,000 year old designs we used the vases best geometric shape as our design intent highlighting tool marks and asymmetries directly on the artifact.
Then it goes into the results of each vase as I linked earlier. Each vase based on the median of the circularity and concentricity and gets a score. The best vase being 0.062mm median error deviation.
www.artifactfoundation.org
But nevertheless I dug out sections where I think this is relevant to the objections. Here Karoyl explains the software used to analyse the vase. Despite the ad hominems about these researchers being amateurs the software developed to handle these unusual vases as precedents in precision metrology had several experts behind it.
Professor Marian Marcis PHD: Photogrametry, image scanning, 3D reconstruction and digitalisation of cultural heritage.
Johannes Bjorn Meyer PHD: Mathmatics, Geometry on negatively curved spaces, Signal Processing, and Medical research Engineering.
Marton Szemenyei PHD: Electrical Engineer, Computer Vision and Deep Learning research, Ai and 3D in Robotic perception.
In fact they are pioneering new software and methods for detecting these unusual and unprecedented precision vases.
From about the 28 minute mark Karoyl Poka who also has a PHD in Electrical Engineering and computer science explains how the software was developed and how it does render 3D models for analysis.
At the 29.20 minute mark it explains the software called Petriescope and how it works in cleaning up images of unnecessary obvious noise.
Segmentation.
Once clean it is exported as a STM Mesh. The realignment is done using Principle Coordinate Analysers (PCA) which analyses all the points in the mesh to find its main orientation of the vase for alignment. From there an automatic segmentation algorythm identilogue is created of all vase sections or points. It only needs two imputs (top and bottom vertices) on the vase and it generates the required vertex groups. This step lets allows testers to isolate intersections while ignoring chips and missing pieces.
Alignment.
Set the vase straight in the grid relative to the global Z axis. The core of the alignment process in Petriescope uses two algorithms. One alogorithm slices the vase in 3D layers as did Maximus.
Each slice is measured for circularity against the best fit circle to the cross section outline of the vase. Then connects the center of all those cross section circles effectively tracing the objects central axis from bottom to top. The estimated axis is then aligned to the global Z axis.
Analysis
Mainly focuses on circularity or how perfectly round a cross section is and concentricity which means how well centered the cross sections are to the global axis Z.
For circularity the software slices align the vase into many horizontal sections almost like a Cat Scan but are a mesh. For each slice it fits a perfect circle to the outline of the cross section. Then it calculates a value called Root Means Square Deviation (RMSD) for that slice. In simple terms RMSD is like an average error. It measures how far each cross section deviates from the perfect circle. Each slice gets its own RMSD value telling us the level of roundness in that slice of the vase.
Then the single metric or (Median) is used for the entire vase as explained earlier. It is the best method as it allows for chips, damage or scanning noise outliers that may skew the findings.
Standard geometric dimensioning and tolerances practice or GD & T often uses a high to low median for roundness. Basically the difference between the maximum and minimum radius in a cross section.
Then placing these circles in a CAD model as the perfect reference points. This produces color coded heat maps that reveal the true surface deviations. This offers a full 3D picture graph view rather than a flat 2D view arrow plot.
Industrial inspection normally compares parts to the original CAD drawings. Since we lack 5,000 year old designs we used the vases best geometric shape as our design intent highlighting tool marks and asymmetries directly on the artifact.
Then it goes into the results of each vase as I linked earlier. Each vase based on the median of the circularity and concentricity and gets a score. The best vase being 0.062mm median error deviation.

Petrie Museum | Artifact Foundation

Last edited:
Upvote
0