Does SAM dream of EIG? Characterizing Interactive Segmenter Performance using Expected Information Gain

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Chung, Kuan-I, Moyer, Daniel
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866910422819405824
author Chung, Kuan-I
Moyer, Daniel
author_facet Chung, Kuan-I
Moyer, Daniel
contents We introduce an assessment procedure for interactive segmentation models. Based on concepts from Bayesian Experimental Design, the procedure measures a model's understanding of point prompts and their correspondence with the desired segmentation mask. We show that Oracle Dice index measurements are insensitive or even misleading in measuring this property. We demonstrate the use of the proposed procedure on three interactive segmentation models and subsets of two large image segmentation datasets.
format Preprint
id arxiv_https___arxiv_org_abs_2404_16155
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Does SAM dream of EIG? Characterizing Interactive Segmenter Performance using Expected Information Gain
Chung, Kuan-I
Moyer, Daniel
Computer Vision and Pattern Recognition
Information Theory
Machine Learning
We introduce an assessment procedure for interactive segmentation models. Based on concepts from Bayesian Experimental Design, the procedure measures a model's understanding of point prompts and their correspondence with the desired segmentation mask. We show that Oracle Dice index measurements are insensitive or even misleading in measuring this property. We demonstrate the use of the proposed procedure on three interactive segmentation models and subsets of two large image segmentation datasets.
title Does SAM dream of EIG? Characterizing Interactive Segmenter Performance using Expected Information Gain
topic Computer Vision and Pattern Recognition
Information Theory
Machine Learning
url https://arxiv.org/abs/2404.16155