Ideal Observer for Segmentation of Dead Leaves

This repository contains the code to apply the Bayesian ideal observer as described in "Ideal Observer for Segmentation of Dead Leaves".

Installation

You can install this included package directly from github through

pip install git+https://github.com/ag-perception-wallis-lab/ideal_observer_dead_leaves_segmentation.git

Dependencies should be installed automatically based on pyproject.toml. The generation of the dead leaves images is done via a separate package (dead-leaves) which is not yet publicly available.

Basic directory structure

• figures: Figures generated for the manuscript.
• ideal_observer: Source code for the ideal observer.
• publications: TeX and pdf files of the paper.
• results: Ideal observer output for example.
• scripts: Demo notebook for application.
• tests: Tests for source code.