Multiscale notes

An emerging effort is to work with multiscale (aka multiresolution, aka pyramid) data arrays. This looks to be a promising component in an overall approach to the visualization of larger-than-memory data.

• Zarr

  • video: Reading multiscale image formats as Zarr with Dask
    • Clip: "we can pass [multiscale] lists of dask arrays to nepari which can lazily load the image pyramid as we interact with the scene"... This would be a game changer for datashader rendering.. it just switches to a different tile/resolution level depending on the screen resolution and zoom level
  • zarr Extension proposal: multiscale arrays
    • continued on image.sc
  • zarr multiscale convention

• OME Zarr - Tools for multi-resolution images stored in Zarr filesets, according to the OME NGFF spec.

  • best practices
  • 
  • image.sc: OME-Zarr storage structure for multiple, multiscale images
    • Allen Institute for Cell Science has an ome zarr writer in aicsimageio has multiscale writing poc

• xarray

  • issue about integrating xarray-datatree into xarray - last activity Feb 2022
    • array datatree blogpost
  • xarray Feature Request: Hierarchical storage and processing in xarray - June 2020
  • xarray-multiscale - scalable multiscale image pyramids in python (by HHMI folks)

• Carbonplan - ndpyramid - A small utility for generating ND array pyramids using Xarray and Zarr.

• Dask working on blogpost

• spatial-image/multiscale-spatial-image - Generate a multiscale, chunked, multi-dimensional spatial image data structure that can serialized to OME-NGFF. Each scale is a scientific Python Xarray spatial-image Dataset, organized into nodes of an Xarray Datatree.