Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Visual artifact with multiscale neuroglancer data #677

Open
LarsKoeppel opened this issue Dec 4, 2024 · 1 comment
Open

Visual artifact with multiscale neuroglancer data #677

LarsKoeppel opened this issue Dec 4, 2024 · 1 comment

Comments

@LarsKoeppel
Copy link

Hi,
I'm experiencing some strange visual artifacts when viewing multiscale Neuroglancer data, and I'm trying to determine whether this issue originates from our data or from Neuroglancer itself.

Here is an excample of this issue

Issue Details:

  • When viewing the data at certain zoom levels, visual artifacts appear on the right side of the image.
  • If you zoom in further, these artifacts disappear, suggesting the issue is related to specific resolution levels.
  • The data is multiscale and precomputed, served via the Neuroglancer precomputed format.

Any insights or suggestions would be greatly appreciated!
@jbms
Copy link
Collaborator

jbms commented Dec 14, 2024

This is actually a tensorstore bug --- will be pushing out a fix soon.

copybara-service bot pushed a commit to google/tensorstore that referenced this issue Dec 20, 2024
@jbms @copybara-github
Fix neuroglancer_precomputed image encoding issues
a5a9613 
Previously, the neuroglancer_precomputed driver incorrectly encoded jpeg
and png-format chunks in the case that the x and y dimensions of the
chunk size differ. In particular, the chunks were encoded with the y
dimension specified as the width, and x * z specified as the height, but
in fact the data was stored with the x dimension as the
inner-most dimension.  While all pixels were stored in the
correct linear order, and could be decoded properly by both tensorstore
and neuroglancer, the image data was misaligned to image rows, such that
individual chunks did not display correctly in a normal image viewer,
and compression performed poorly due to this misaligned.  While png is
lossless and poor compression merely increases the size, with jpeg this
also tended to introduce extreme artifacts.

This commit also enables support for 16-bit PNG images with up to 4
channels.

Fixes google/neuroglancer#677.

PiperOrigin-RevId: 708415255
Change-Id: I057cc3fc4f073026e8bf96cc543be5a769c626c2
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@jbms @LarsKoeppel