Jump to content


  • Content Count

  • Joined

  • Last visited

  • Days Won


mikefazz last won the day on October 10 2017

mikefazz had the most liked content!


About mikefazz

  • Rank
    Power Contributor
  • Birthday 07/28/1980

Recent Profile Visitors

1,592 profile views
  1. While I am not a fan of thresholding for segmentation, a decent scan should be able to distinguish between cortical and trabecular bone with different thresholding parameters. You would need to run thresholding 2 separate times with different thresholding ranges. My preferred method involves using region growing, where you could seed the trabecular and cortical bone with separate seeds and get separate entities.
  2. Nice review and welcome to the joys of desktop FDM printers;) I'm a bit surprised they don't use a PEI coating on the glass bed, I have found it to help with bed adhesion. Another thing to try is using some PVA dissolved in water (fairly saturated) to coat the bed then heating the bed to leave a 'skin' for printing onto. I'm pretty sure the glue stick is basically PVA. Also a wider and thicker first layer will help with bed adhesion this can be adjusted in CURA's advanced print settings. For the clogged nozzle I have recently started doing 'cold' pulls (really warm pulls) with some old nylon I have heating it to ~240C then letting it cool to ~130C and pulling. I do a steady pull instead of jerking and hold the frame near the hot end so as to not put force on the entire machine. I kind of cringed when I saw how hard you pulled. I have been surprised to see just how much black junk was in some of my nozzles. The successful prints do look good, I think they have a better formulation of PVA than I currently use, unfortunately they use 3mm filament and I am all 1.75mm.
  3. Hi Peter, I don't mind taking a look at the data though if it is from a clinical scan it may not be of the correct format or type to create a 3D model from. To get a 3D model I need a 3D volume of data, the other US scan I have looked at was just a random collection of images, some renderings, some montages. As a first step open the data in ImageJ to see what you have if you are not sure. Also I wouldn't think of 3DSlicer as a 'conversion' software. Taking a volume of data and getting a desired 3D (surface) model is usually more of a process. Mike
  4. Mauricio I would be interested in taking a look do you have a vol file I can look at? In general 3D US renderings from scans look to have a long way to go before it is anywhere close to the quality of a decent MRI or CT Mike
  5. mikefazz

    Teeth Micro CT

    I haven't tried importing tiff's but one thing to try would be to use ImageJ to convert a stack of tiffs to an nrrd file for loading in 3DSlicer. Mike (not Dr. Mike;)
  6. Hi I am assuming you have a scan from a gated MRI or a CT scan. If so I would try region growing as shown in my MRI segmentation tutorial:
  7. Hi I don't have any experience with 3D US but would be interested in looking at the data. Do you have dicom files or do they work with a proprietary file type?
  8. Very nice shells... but its been done even further back https://www.themarysue.com/3d-printed-crab-shells/
  9. I printed my hand a couple weeks back. The model is available for sale at: Or try your 'hand' at segmenting it from the scan data that is free at: I am still working on getting better transparency to show the internal bones. With FDM true transparency only works for single perimeter prints (like vases) but I am trying some other plastics that should do better than this one done with PLA. The light source is pretty bright, the bones are difficult to see normally.
  10. Depending on what access you have to scanners and the time you want to invest it is possible to use both a CT and an MRI scan to get good bone as well as soft tissue models. The two scans can be registered together similar to what I showed in my registration tutorial: Also MRI's can be segmented if the quality is high enough. You want your scan resolution to be as high as possible. Ideally the smallest anatomy would have multiple voxels across it. So if you have a tendon that is 4mm across you want an in-plane and out of plane resolution of at least 2mm. High res MRI's take more time which can add cost and require the subject to stay still longer. Also if possible work with the MRI tech to highlight the anatomy of interest; the settings to get good fat tissue will differ from getting good tendons. Segmentation of high quality MRI If MRI is not an option (typically much more $$) then CT will work great for bones but soft tissue will be difficult. Of course if you want some help with segmenting you can contact me through my business www.med-mod.com
  11. I've looked a little into FDA approved materials and am interested in what it takes for a material to be considered safe for surgical guides. The dental material above adheres to the following: EN-ISO 10993-1:2009/AC:2010, EN-ISO 20795-1:2013, EN-ISO 7405:2009/A1:2013 The material guidel!ne from taulman3d (http://taulman3d.com/guidelne-spec.html) meets: ISO 11607-1: 2006, ISO 10993, USP Class VI, USP <661>, DMF (Drug Master File) number 16525 Taulman3d also has nylon 680 (http://taulman3d.com/nylon-680-spec.html) CAS Reg. No. 51995-62-1 meets: 21CFR177.1500 / CFR177.1395 testing follows regulations 21 CFR Parts 210, 211 and 820 Now I realize just because the material is approved doesn't mean it is approved after printed as it has to 'go through' the machine so that may require a COA. Of course any advice here is not taken as 'legal advice' just interested in what it takes to be able to properly produce surgical guides. From my understanding it is mostly a requirement of the client to get FDA approval, I would expect a surgical guide to be at least class 2 and likely class 3.
  12. I haven't worked with a tiff stack but I wonder where it gets the voxel size values from? I would check the tiff meta data and see what values are there. Guessing images are in DPI and fiji reads the values as mm... but then I would think the voxel size would be much smaller. I would expect the same result from ImageJ but worth a try too.
  13. So I have seen some questions here on embodi3D asking how to work with MRI data. I believe the main issue to be with attempting to segment the data using a threshold method. The democratiz3D feature of the website simplifies the segmentation process but as far as I can tell relies on thresholding which can work somewhat well for CT scans but for MRI is almost certain to fail. Using 3DSlicer I show the advantage of using a region growing method (FastGrowCut) vs threshold. The scan I am using is of a middle aged woman's foot available here The scan was optimized for segmenting bone and was performed on a 1.5T scanner. While a patient doesn't really have control of scan settings if you are a physician or researcher who does; picking the right settings is critical. Some of these different settings can be found on one of Dr. Mike's blog entries. For comparison purposes I first showed the kind of results achievable when segmenting an MRI using thresholds. With the goal of separating the bones out the result is obviously pretty worthless. To get the bones out of that resultant clump would take a ridiculous amount of effort in blender or similar software: If you read a previous blog entry of mine on using a region growing method I really don't like using thresholding for segmenting anatomy. So once again using a region growing method (FastGrowCut in this case) allows decent results even from an MRI scan. Now this was a relatively quick and rough segmentation of just the hindfoot but already it is much closer to having bones that could be printed. A further step of label map smoothing can further improve the rough results. The above shows just the calcaneous volume smoothed with its associated surface generated. Now I had done a more proper segmentation of this foot in the past where I spent more time to get the below result If the volume above is smoothed (in my case I used some of my matlab code) I can get the below result. Which looks much better. Segmenting a CT scan will still give better results for bone as the cortical bone doesn't show up well in MRI's (why the metatarsals and phalanges get a bit skinny), but CT scans are not always an option. So if you have been trying to segment an MRI scan and only get a messy clump I would encourage you to try a method a bit more modern than thresholding. However, keep in mind there are limits to what can be done with bad data. If the image is really noisy, has large voxels, or is optimized for the wrong type of anatomy there may be no way to get the results you want.
  • Create New...