FREEDOM251

Freedom251 › Compare

Side by side

LabelMe vs MakeSense.ai

Both stand on ai image annotation tools. Which one you can start free with, which one lets you read the code, and what each costs once you pay.

LabelMe

Free plan LinuxSelf-hostedWeb+1

LabelMe provides annotation modes, collection navigation and XML output for image-labeling work. Its web-server requirements and file constraints are important to review before setup.

Visit LabelMe

MakeSense.ai

Free plan Open source LinuxMacSelf-hosted+2

Pick MakeSense.ai if you need a free browser tool for photo annotation and dataset preparation, especially for a small computer vision project. Check that its format matrix supports your annotation and import/export needs before choosing it, since some combinations are not yet available.

Visit MakeSense.ai

The facts

LabelMeMakeSense.ai
Free planYesYes
Open sourceNot publishedYes
Runs onLinux, Self-hosted, Web, WindowsLinux, Mac, Self-hosted, Web, Windows
Lowest paid planNot publishedFree or on request
Plans publishedNone1
MakerNot publishedNot published
Websitelabelme.csail.mit.edumakesense.ai
On ai image annotation tools#7 of 24#3 of 24

LabelMe: what it does well

  • Free to use.
  • Supports polygon, scribble and mask annotations.
  • Exports annotations as XML.
  • Includes a MATLAB toolbox and tar-file downloads.
  • URL options support Mechanical Turk workflows.

What to know first

  • Installation requires Apache and Perl/CGI support.
  • Quick start specifies .jpg image files.
  • Online dataset access may be slow.
  • Server folders require specific read and write permissions.

MakeSense.ai: what it does well

  • Runs in a browser across operating systems.
  • Supports five annotation types, including polygons and labels.
  • Offers listed AI integrations including YOLOv5, SSD and PoseNet.
  • Images are not sent to a server, according to the project.
  • Licensed under GPL-3.0.

What to know first

  • Some annotation and import/export format combinations are not yet supported.
  • Custom models must be YOLOv5 exports in TensorFlow.js format.

Alternatives to LabelMe Alternatives to MakeSense.ai The whole list