Atlas Maker

Client - Neurome, Inc.

 

(Circa 2004) Atlas Maker is a Macintosh and Windows application designed to facilitate the organization of brain serial sections acquired by either its internally supported video and scanner procedures or from external datasets created by other applications and the creation of a brain atlas. The application is designed to be very easy to use, similar to Apple's iPhoto, presenting rapidly displayed thumbnails with full size views. A full vector based contouring system is also programmed into Atlas Maker so that neuronal structures may be outlined with Bezier tools. Importing and exporting of data are XML based.

 

Workflow

 

The general workflow for creating a brain atlas is displayed here. The basic steps are:

  • Datasets
  • Thumbnails
  • Sections
  • Classification
  • Normalization
  • Comparison

Datasets

 

  • Display any number of NeuroMosaic derived (or other) image sets
  • Import from any network location
  • Export to other file formats
  • Create from attached video (i.e., microscope camera, EM, MRI scanner?) or scanner (TWAIN)
  • Identify attributes (species, strain, organ, stain) and progress phase of modeling workflow

Thumbnails

 

  • Quick and easy to navigate overview of everything file in the dataset (Macintosh iPhoto inspired)
  • Thumbnails created as is needed and automatically
  • New parameters for thumbnails can be specified - resolution detail, size

Sections

 

  • Display one section at a time from dataset
  • Authoring tools - selection, grab, polyline, Bezier, vertex, rectangle, oval, copy contours, dilate, contract, zero, scale up, scale down, navigation, transparency (alpha)…
  • All data are written as it is entered
  • All XML based

Classification

 

  • The unnamed contours are processed automatically to create ROI masks that are used to extract the fully detailed pixmap information from the image datasets
  • Detailed ROIs are displayed by authoring tool sequentially for user identification (pick list, speech recognition, etc)
  • Named ROIs are stored in dataset
  • ROIs are intrinsically 2D (pixmaps)