BigDataViewer-based tool for browsing multichannel multiscale N5 datasets.

The following packages are required for installation:

• Apache Maven - https://maven.apache.org/
• Oracle JDK / OpenJDK - https://www.oracle.com/technetwork/java/javase/downloads/index.html

To install the application as a Fiji/ImageJ plugin, run the following:

python install.py <path to Fiji.app>

Then, it will be available in Fiji under Plugins -> BigDataViewer -> N5 Viewer.

The plugin supports multiple storage options:

• Filesystem (local/network drives)
• Amazon Web Services S3
• Google Cloud Storage

Datasets stored on the filesystem are represented by the root N5 directory. For cloud storages, every N5 container is a bucket. You will be prompted to select an appropriate storage type and an N5 container as an input for the plugin.

The plugin specifies the following N5 container structure:

/attributes.json

/c0/attributes.json
/c0/s0/
/c0/s1/
/c0/s2/
/c0/...

/c1/attributes.json
/c1/s0/
/c1/s1/
/c1/s2/
/c1/...

...

Root attributes are used as defaults for all channels. They can be overridden by setting channel-specific attributes.
s0, s1, s2 are standard N5 datasets representing scale levels. s0 corresponds to full resolution.

Example of the attributes.json file:

{
  "name":"some data",
  "scales":[[1,1,1],[2,2,1],[4,4,2],[8,8,4],[16,16,9],[32,32,17]],
  "pixelResolution":{"unit":"um","dimensions":[0.097,0.097,0.18]},
  "affineTransform":[[1,-0.30,-0.25,0],[0,1.25,0,0],[0,0,0.85,0]]
}

Alternatively, scales and pixel resolution can be stored and queried as attributes of individual scale level datasets. The attribute keys for individual values are downsamplingFactors and pixelResolution. Thus, root attributes or channel group attributes are not mandatory, which makes it fully compatible with scale pyramid exports generated with N5 Spark.

Fetching data from cloud storages requires authentication. You will be prompted for your user credentials on first use of the respective cloud service. For Google Cloud, a browser tab will pop up with web-based authentication prompt. For AWS, the credentials need to be initialized with aws configure, more information is available here.

The changes made to the viewer state are saved automatically to bdv-settings.xml file in the root N5 container:

• brightness & contrast
• color scheme
• display mode (fused/single, interpolation)
• channel grouping
• channel transformations
• bookmarks

To prevent concurrent modification, exclusive file locking is enforced, and the application will warn and suggest to open in read-only mode if somebody is already browsing the dataset.

The application has a built-in cropping tool for extracing parts of the dataset as a ImageJ image (can be converted to commonly supported formats such as TIFF series).

Place the mouse pointer at the desired center position of the extracted image and hit SPACE. The dialog will pop up where you can specify the dimensions of the extracted image.

The image is cropped without respect to the camera orientation, so the cropped image will always contain Z-slices.