Jump to Content
Centaur Labs Reference Documents
HomeGuidesRecipesAPI ReferenceChangelogDiscussions
v1.0

Log InCentaur Labs Reference Documents
Guides
Log In
v1.0HomeGuidesRecipesAPI ReferenceChangelogDiscussions
All
Pages
Start typing to search…

Getting Started

  • Introduction
  • Labeling interfaces
  • Quick start guide
  • Key definitions

How To Guides

  • Account set up
    • Billing
    • Setup and invitations
    • Get support
    • Security
    • Workspace
  • Create project and import data
    • Import Data
      • Import data with S3
      • Self-host data on the cloud
      • Import Data from Presigned URL
    • Define project views
    • Create files manifest
    • Supported data types
      • Format HTML data
      • DICOM Best Practices
    • Manage dataset
    • Use tags
      • Organize data using tags via API
  • Define labeling task
    • Manage task
  • Create Prelabels
  • Create Gold Standards
  • Create instructions
    • Example 1
    • Example 2
  • Labeling
    • DiagnosUs
    • Web Labeling
    • Centaur DICOM Viewer 101
      • Troubleshooting DICOM Viewer
      • Slice-to-Slice Annotation Duplication
  • Download and analyze results
    • Download results
    • How to review results
    • Improve future results
    • Thresholding

Conceptual Guides

  • Labeling states
    • Label assignment
    • Performance measurement
  • Label collection
    • Qualified Reads
  • Label aggregation
    • Classification
    • Polygons and Pixel
    • Boxes, lines, circles
    • Range Selection
  • Case level metrics
    • Classification
    • Polygons and Pixel
    • Boxes, lines, circles
    • Range Selection
  • Task level metrics
    • Classification
    • Segmentation
    • Range Selection
    • Multi-Class Confusion Matrices
  • Label export glossaries
    • Classification
    • Polygons
    • Pixel
    • Cuboid
    • Boxes, lines, circles
    • Ranges
    • WKT Format
    • API JSON Results

LABELER GUIDES

  • Payments
Powered by 

Label aggregation

Once reads are collected, they are aggregated to provide the optimal Majority Label. Aggregation methodology varies based on the annotation type - click on the links below for additional information.

  • Classification
  • Polygon and pixel segmentation
  • Boxes, lines or circles segmentation
  • Range selection

Updated 3 months ago


Qualified Reads
Classification