/ Machine Learning

Design your own smart Image Annotation process in a web-based environment

Design your own smart Image Annotation process in a web-based environment


Label Objects and Save Time (LOST) - Design your own smart Image Annotation process in a web-based environment.


LOST (Label Object and Save Time) is a flexible web-based framework for semi-automatic image annotation. It provides multiple annotation interfaces for fast image annotation.

LOST is flexible since it allows to run user defined annotation pipelines where different annotation interfaces/ tools and algorithms can be combined in one process.

It is web-based since the whole annotation process is visualized in your browser. You can quickly setup LOST with docker on your local machine or run it on a web server to make an annotation process available to your annotators around the world. LOST allows to organize label trees, to monitor the state of an annotation process and to do annotations inside the browser.

LOST was especially designed to model semi-automatic annotation pipelines to speed up the annotation process. Such a semi-automatic can be achieved by using AI generated annotation proposals that are presented to an annotator inside the annotation tool.

Getting Started

If you feel LOST,
please find our full documentation here: https://lost.readthedocs.io.

LOST QuickSetup

LOST releases are hosted on DockerHub and shipped in Containers.
For a quick setup perform the following steps (these steps have been
tested for Ubuntu):

  1. Install docker on your machine or server:

  2. Install docker-compose:

  3. Clone LOST:

    git clone https://github.com/l3p-cv/lost.git
  4. Run quick_setup script:

    cd lost/docker/quick_setup/
    # python3 quick_setup.py path/to/install/lost
    python3 quick_setup.py ~/lost
  5. Run LOST:

    Follow instructions of the quick_setup script,
    printed in the command line.

Citing LOST

    title={{LOST}: A flexible framework for semi-automatic image annotation},
    author={Jonas J\"ager and Gereon Reus and Joachim Denzler and Viviane Wolff and Klaus Fricke-Neuderth},
    Journal = {arXiv preprint arXiv:1910.07486},