I have a series of 75 images similar to the one attached. The images go all around the anatomy and logically should make an easy mesh due to the detail and clear contrast. Meshroom will only green check a couple in the series and absolutely nothing in terms of creating a mesh. Thoughts on settings to make this work? Thanks.

Hi,
First, you could try to change the preset on the FeatureExtraction from normal to high.
If that doesn't solve your issue, you could send us multiple images to test it.
You could use fromsmash and our private mailing-list [email protected].
Best,
I am certainly not qualified to provide an answer, but here's a question....if all of the images have the same scale overlay, and alphanumeric info, won't the software try to align the photos on those features as well? I'm guessing that a set of images without that overlay would probably work
@jamesmoe Did you manage to get decent results?
Nope. Took out all the lettering in the background. Turned settings all the way up. There is no reason this should not have worked. I ended up using something called 3D slicer, but that was a pain cause I had to basically learn how to read a ct scan to isolate the anatomy.
On Feb 11, 2019, at 6:46 AM, Fabien Castan <[email protected]notifications@github.com> wrote:
@jamesmoehttps://github.com/jamesmoe Did you manage to get decent results?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHubhttps://github.com/alicevision/meshroom/issues/380#issuecomment-462353125, or mute the threadhttps://github.com/notifications/unsubscribe-auth/ABid5C9J-3qqDAnBk07lBZe2wa0oujrvks5vMYJBgaJpZM4abE59.
Any chance you can send me the images privately to understand why it's not working?
You bet. Have a Dropbox or something I can zip them up and send?
On Feb 11, 2019, at 10:34 AM, Fabien Castan <[email protected]notifications@github.com> wrote:
Any chance you can send me the images privately to understand why it's not working?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHubhttps://github.com/alicevision/meshroom/issues/380#issuecomment-462441447, or mute the threadhttps://github.com/notifications/unsubscribe-auth/ABid5N8BVdfFBAG1LyiNZN_tgw8DqlMxks5vMbepgaJpZM4abE59.
just out of curiosity, i am wondering how much overlap you have for adjacent images? and how much difference in camera angle?
@jamesmoe Could you send me an email to [email protected] to find a way for sending files?
FYI for the developers and others on this particular data set.
The source data is a CTA study of the aortic arch, and it was a 3D model to begin with. The rendered images were artificially created with fake lighting and rendering. So, the question really is, what is the goal of meshing this? If the goal is a 3D rendering of the aortic arch, then just take the original data, run it through slicer or itk-snap, then apply a texture to the output. If for some reason you are trying to recreate the 3D source data, I guess you could try to mesh this, but presumably it is stored somewhere.
I ended up learning enough of 3D Slicer to isolate the aorta from the source CTA and export it to STL.
It was a bit of a learning curve for what I was trying to accomplish.
It would have been much easier to take this and run it through meshroom treating it like photos around a 3d object. It should totally have worked as it is much cleaner than a real series of images out in the wild.
Sent from Mailhttps://go.microsoft.com/fwlink/?LinkId=550986 for Windows 10
From: RadCLarite notifications@github.com
Sent: Monday, March 18, 2019 7:48:14 PM
To: alicevision/meshroom
Cc: jamesmoe; Mention
Subject: Re: [alicevision/meshroom] Uniform medical images will not mesh (#380)
FYI for the developers and others on this particular data set.
The source data is a CTA study of the aortic arch, and it was a 3D model to begin with. The rendered images were artificially created with fake lighting and rendering. So, the question really is, what is the goal of meshing this? If the goal is a 3D rendering of the aortic arch, then just take the original data, run it through slicer or itk-snap, then apply a texture to the output. If for some reason you are trying to recreate the 3D source data, I guess you could try to mesh this, but presumably it is stored somewhere.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHubhttps://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Falicevision%2Fmeshroom%2Fissues%2F380%23issuecomment-474144435&data=02%7C01%7C%7C40d515456b334b57826008d6abfc3065%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C636885496958395741&sdata=2PCmTKj8tafpPVtIfavxAHw%2B2nH7%2FnIDaZn6ue9izQ8%3D&reserved=0, or mute the threadhttps://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FABid5OUE20QSpMvUvKKG4-vUG2LBrLNjks5vYCW-gaJpZM4abE59&data=02%7C01%7C%7C40d515456b334b57826008d6abfc3065%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C636885496958405740&sdata=IqI079vEMSlYcQPh1LiAZFqUftoM1JsSLYgBCmo%2FuTo%3D&reserved=0.
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
This issue is closed due to inactivity. Feel free to re-open if new information is available.
Most helpful comment
I am certainly not qualified to provide an answer, but here's a question....if all of the images have the same scale overlay, and alphanumeric info, won't the software try to align the photos on those features as well? I'm guessing that a set of images without that overlay would probably work