Table Of Contents

Note: This document and those linked to it contain the beginnings of the user documentation for BioImage Suite Web. A brief introduction to the software can be found in this presentation. If you are looking for developer documentation, this may be found in the doc directory of the source repository.

This document represents work in progress.


Starting BioImage Suite Web

BioImage Suite Web consists of three major components

Introduction Page This is the main page at https://bioimagesuiteweb.github.io/webapp/

BioImage Suite Web consists of a collection of applications (which is likely to grow over time.) The list can be accessed under the Applications menu (as shown) and snapshots and brief descriptions of each application can be seen in the carousel (slide show) at the top of the page.

From the menu you can also navigate to the downloads page, the documentation and the source code repository.


Some Key Information

Download File Location

When using BioImage Suite Web in a web browser all save operations are implemented as download operations, i.e the software creates the data output and initiates a download event. By default most browsers send all downloads to your Downloads directory which is completely unsuitable for using BioImage Suite Web as an analysis software. We recommend going under Settings in your favorite browser as shown below for Chrome:

Chrome Download Settings

You need to turn the option Ask where to save each file before downloading to ON.

For Safari on MacOS see the description on this web page.

Image Orientations

The Viewers and commandline tools can load images in the NIFTI-1 .nii.gz format. A key component of the image is the image orientation. This relates the internal image coordinates to “world” coordinate space.

Image Orientation

If we consider a 3-dimensional image matrix F(i,j,k) where i,j and k are the coordinate axes in the images where i,j are the in-plane directions and k is the slice axis. For example in the image shown above for Axial images, i is often Left->Right, j is Posterior->Anterior. The slice axis k is often Inferior->Superior. Such an image is described as having orientation RAS where the letters stand for the direction of the individual axes i.e

RAS is probably the most common orientation for Neuroimaging research. An alternative orientation that is commonly used (and which the legacy BioImage Suite used) is LPS which is 180 degrees rotated w.r.t. RAS (flip i and flip j).

There are 48 (!) possible combinations of orientations. BioImage Suite Web can display all of these (at least as far as we have tested) but also provides the user with the option of converting images on-load to RAS or LPS to standardize. Under the Help menu if you select the option Set Image Orientation on Load the following GUI appears.

Image Orientation on Load

If you select either RAS or LPS then any image load will be repermuted to be in RAS or LPS orientation on load. The settings are stored either in the Browser Database or the text file ${HOME}/.bisweb for commandline and Desktop applications. This is a JSON key-value database file that may look something likes this:

The Settings File

In this case “orientationOnLoad” is set to “None” which preserves the image orientation as is.


The Orthogonal Viewer Application

All BioImage Suite web applications share components and have a similar user interface. The picture below shows the Orthogonal Viewer Tool. This viewer consists of a single viewer plus some controls. All viewers in BioImage Suite Web can display two images:

  1. The underlay “image” – this is often the anatomical image and be loaded and saved from the Image Menu.
  2. The “overlay” or “objectmap” image – this is often a functional image or a segmentation image. This can be loaded and saved from the Overlay menu.

An Application

The application consists of the following parts:


The Overlay Viewer Application

All viewers can display overlays. The Overlay Viewer is particularly optimized for this task. This is shown below.

The Overlay Viewer

This is a dual viewer that shows functional overlays in both an Orthogonal (three orthogonal slice viewer shown above) and a Mosaic (multiple parallel slices – see below) viewer which are synchronized to have the same color mapping scheme.

To try this out, open the application from the main BioImage Suite web page (under Applications| Overlay Viewer) and then under Help, select Load Sample Data. This will load a sample anatomical/functional image pair. (You may load your own images using Image|Load Image and Overlay|Load Overlay. As discussed above, once an overlay image is loaded, the Viewer Controls have an additional set of controls labeled Overlay Color Mapping (shown on the right). The controls here perform the following tasks:

This viewer can also be switched to Mosaic view by clicking the Mosaic Tab (just below the menu).

Mosaic Mode for Overlay Viewer

This shows multiple parallel slices of a particular orientation (selected using Plane). You may set the number of rows and columns, the first slice and increment (which may be negative). The colormapping controls are identical to the orthogonal (3-slice view) and is synchronized between the Orthogonal and Mosaic tabs.


Image Editor

Introduction

The Image Editor Tool

The Image Editor tool shown above is a tool to create and visualize interactive segmentations, as well as to correct existing segmentations (e.g. skull stripping). This tool shares much of the functionality of the The Orthogonal Viewer Application with the following additions:

Creating an Objectmap

Threshold Image

One can create an objectmap on one of three ways:

Detailed Operations

Image Editor Parts

Once an objectmap is n memory it can displayed and manipulated using tools provided in the Image Editor Tool. The figure above highlights five different pieces of functionality as follows:

The following is a brief description of the functionality included in this:

C. The Morphology Operations Tool – this performs mathematical operations such as erode, dilate, median, connect (seed connectivity from current cross-hairs) etc on 0/1 objectmaps. If the objectmap contains multiple colors then they will all be converted to 0 or 1.

D. The Regularize Objectmap Tool – this magic tool performs markov random field regularization to smooth manually painted regions. It was developed for the construction of the Yale Brodmann atlas image. (You can load this under Obejctmap Load Yale Brodmann Atlas.) To get a desired level of smoothness, set the smoothness value and press Smooth.

E. The Mask Image Tool – this can be used to mask the underlying anatomical image with the objectmap to mask out all parts of the image outside the mask (if Inverse is set to Off ). Setting Inverse to On allows one to mask the region inside the objectmap. Before masking, the objectmap must be binarized by thresholding it using the threshold set using the Threshold slider.

VOI Analysis

Given an objectmap and a 4D image we can generate image timeseries plots as shown above. To invoke this control go to ‘Objectmap VOI Analysis’. By default, the graph shows the average intensity in each region over time. There are five buttons at the bottom of this which perform the following operations:

This page is part of BioImage Suite Web. We gratefully acknowledge support from the NIH Brain Initiative under grant R24 MH114805 (Papademetris X. and Scheinost D. PIs, Dept. of Radiology and Biomedical Imaging, Yale School of Medicine.)