Are MSER features really interesting?

Ron Kimmel; Cuiping Zhang; Alex Bronstein; Michael Bronstein

doi:10.1109/TPAMI.2011.133

Are MSER features really interesting?

Ron Kimmel, Cuiping Zhang, Alex Bronstein, Michael Bronstein

Computer Science

Technion - Israel Institute of Technology

Research output: Contribution to journal › Article › peer-review

Abstract

Detection and description of affine-invariant features is a cornerstone component in numerous computer vision applications. In this note, we analyze the notion of maximally stable extremal regions (MSERs) through the prism of the curvature scale space, and conclude that in its original definition, MSER prefers regular (round) regions. Arguing that interesting features in natural images usually have irregular shapes, we propose alternative definitions of MSER which are free of this bias, yet maintain their invariance properties.

Original language	English
Article number	5936074
Pages (from-to)	2316-2320
Number of pages	5
Journal	IEEE Transactions on Pattern Analysis and Machine Intelligence
Volume	33
Issue number	11
DOIs	https://doi.org/10.1109/TPAMI.2011.133
State	Published - 2011

Keywords

MSER
affine invariance
correspondence
feature detector
stable region

All Science Journal Classification (ASJC) codes

Software
Computer Vision and Pattern Recognition
Computational Theory and Mathematics
Artificial Intelligence
Applied Mathematics

Access to Document

10.1109/TPAMI.2011.133

Cite this

@article{a59aa791e6304148a8cc1857686e7c4f,

title = "Are MSER features really interesting?",

abstract = "Detection and description of affine-invariant features is a cornerstone component in numerous computer vision applications. In this note, we analyze the notion of maximally stable extremal regions (MSERs) through the prism of the curvature scale space, and conclude that in its original definition, MSER prefers regular (round) regions. Arguing that interesting features in natural images usually have irregular shapes, we propose alternative definitions of MSER which are free of this bias, yet maintain their invariance properties.",

keywords = "MSER, affine invariance, correspondence, feature detector, stable region",

author = "Ron Kimmel and Cuiping Zhang and Alex Bronstein and Michael Bronstein",

note = "Funding Information: This research was supported by the European Commission{\textquoteright}s FP7-ERC program, grant agreement no. 267414. Michael M. Bronstein is supported by the Swiss High-Performance and High-Productivity Computing (HP2C) grant.",

year = "2011",

doi = "10.1109/TPAMI.2011.133",

language = "الإنجليزيّة",

volume = "33",

pages = "2316--2320",

journal = "IEEE Transactions on Pattern Analysis and Machine Intelligence",

issn = "0162-8828",