Dear Lam Dao - Ngoc,

Pls check the link and the attachment.

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4209425/

Appreciated to your support, Professor Nirmala.

Regards.

Dear all,

Thank to all support from Professor Nirmala.

I would like to clarify my question a little bit for better discussion and supports.

The research on the engineered accuracy of 3D CT imaging devices is one of the lacking in dental studies, due to the unavailability of devices suppliers sensitivity. Factually, most of those studies in this field just interested in the imaging output results only, via indirect method through the 3D physical measurement. That is why I set the question here.

In my current open project on the another approach of digitalized dental implant surgery guided splint/ template issue protocol, I try to identify the real magnification and imaging deformity while converting from DICOM data set (called object A) into the 3D STL model (called object A'), before doing any planning and registration with the 3D optic scanner on the virtual and physical cast model. The checking will be paired evaluated. I will used that factor to correct the matching between the 3D STL model from DICOM dataset and the 3D STL from 3D optic scanner before making the registration. That will make better result for the final rapid prototyping model fitness and accuracy.

I jut designed and manufactured a calibration template (cubic template, machined on 4 axis milling machine with the accuracy of +/- 0.05 mm ) for identify the following factors on the 3D CT imaging devices, such as:

1. the X/Y/Z axis magnification

2. the /Y/Z/X axis imaging deformity (around it)

3. the smoothness of reconstruction results on RAW dataset on the device software

4. the optimum marking technique for the registration

I am planning to make my project on the following 3D CT devices in Ho Chi Minh City, Vietnam, such as:

1. Picasso (Vatech, Korea) (done)

2. Pax-i Smart (Vatech, Korea) (done)

3. RaySCAN Symphony BM Option (Ray Medical, Korea)

4. RayAlpha (Ray Medical, Korea)

5. HDX (HDX, Korea)

6. Galileos Scanner (Sirona, Germany)

7. Orthophos (upgraded) (Sirona, Germany)

8. GDX-300 (Gendex) (done)

9. Accuitomo (Morita, Japan)

All the STL data will be compared on meshing compare software call ccloud.

The reference photos of the calibration template and some step on Pax-i Smart (Vatech) is attached in my project at: https://www.researchgate.net/project/Dental-implant-surgery-guided-splint-freeware-solution_57296170f7b67ef9ec506387

In the other hand, why i would like to do this study? Because I would like to double-check my thinking only.

After reading the researches in accuracy of stereolithography dental implant susrgery guided splint/template, i think the accurcay of that might be have the another issues. In my opinion, the basics of the final accuracy is based on the following issue:

1. the ‘fitness’ of power of X-ray imaging tube,

2. the X-ray CT imaging system alignment (including the colimator and imaging sensor positioning relation)

3. the stability/rigidity of C-arm of Dental 3D CT imaging devices (in termed of ‘systematic imanging mechanical vibration offsetting’)

4. the proper matching between the STL models generated from DICOM dataset and 3D optic scanned data

5. the accuracy of 3D optical scanner (mostly in the range of +/- 10 microns)

6. the optimum fitting of the STL files before planning

7. the accesibility or the comfortability of implantologist while making the implant placing with the surgery guided splint/ template, because of prevention of the displacement of splint/ template while doing the surgery

8. the accuracy of 3D printer (just mention on the DLP printer with the accuracy in the range of +/- (25 - 100) microns)

9. the fitness between the guiding sleeves and the splint’s/ template’s boring sleeve stocked holes (including the accuracy of the final pritned splint/template, the co-centricity of the sleeve’s hole and the sleeve’s outer diameter (most of machined tollerance of milled objects is around +/- (0.02 - 0.05) mm and the bonding gap)

10. the post-processing (after making the splint/template and before implant placing) and post-operation (after implant placing). in my oppinion, please do not use the X-ray imaging control, due to its value is not adapted to the final accuracy needs.

And other 5 other issues that i will not mention here. Two of them are about the method how to double-checke the accuracy of the drilling guided direction and the post-op placed implant direction in the indirect way. :) Another five are mention on the follong issues, such as: 1. - patient reaction while operation, 2. - the elastic deformity of splint/ template, 3. - the systematic assembly adaption/ tollerance between drilling bits, stoppers and guiding sleeves.

Looking forward to your support about the relevant studies and advices.

Regards,

Dao - Lam

Puhhh - this is just to much! Seriously, I recommend to seperate your problem into some parts and open new questions about it.

What I understood so far is, that you want to make a comparison of two (or more) different techniques with a meshed 3D model. In the field of photogrammetry a comparison of different 3D models is usually done like this:

• Define a set of control points. These control points should be well distributed over the model.
• These control points have to be identified in both 3D models.
• Use these control points in order to do a transformation from one to the other model. This is called registration, as you certainly already know.
• Such a registration incorporates a 3D-similarity transformation (or Helmert-transformation). Because it (normally) uses more control points than necessary it is done with least-squares-adjustment. The outcome of such an adjustment is not only the set of transformation parameters, but also the precision (standard deviation) of how well the control points fit together between both models
• after the registration you pick some more so called check points at different positions off the control points. Transform these check points and compute the pairwise distance of these check points between both models. From this you can calculate the root mean square (RMS) of these distances. The RMS is compareable to the standard deviation of the control points from the adjustment (during registration). In other words, this gives you a measure of how well both models fit together.

I hope it helps!

Regards

Olaf

Perhaps this article about a new intraoral scanner might be helpful

Optics and Lasers in Engineering

Volume 54, March 2014, Pages 187-196