Is multi-model feature matching better for endoscopic motion estimation?

Xiang Xiang, Daniel Mirota, Austin Reiter, Gregory Hager

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

Camera motion estimation is a standard yet critical step to endoscopic visualization. It is affected by the variation of locations and correspondences of features detected in 2D images. Feature detectors and descriptors vary, though one of the most widely used remains SIFT. Practitioners usually also adopt its feature matching strategy, which defines inliers as the feature pairs subjecting to a global affine transformation. However, for endoscopic videos, we are curious if it is more suitable to cluster features into multiple groups. We can still enforce the same transformation as in SIFT within each group. Such a multi-model idea has been recently examined in the Multi-Affine work, which outperforms Lowe’s SIFT in terms of re-projection error on minimally invasive endoscopic images with manually labelled ground-truth matches of SIFT features. Since their difference lies in matching, the accuracy gain of estimated motion is attributed to the holistic Multi-Affine feature matching algorithm. But, more concretely, the matching criterion and point searching can be the same as those built in SIFT. We argue that the real variation is only the motion model verification. We either enforce a single global motion model or employ a group of multiple local ones. In this paper, we investigate how sensitive the estimated motion is affected by the number of motion models assumed in feature matching. While the sensitivity can be analytically evaluated, we present an empirical analysis in a leaving-one-out cross validation setting without requiring labels of ground-truth matches. Then, the sensitivity is characterized by the variance of a sequence of motion estimates. We present a series of quantitative comparison such as accuracy and variance between Multi-Affine motion models and the global affine model.

Original languageEnglish (US)
Title of host publicationComputer-Assisted and Robotic Endoscopy - 1st International Workshop, CARE 2014 held in Conjunction with MICCAI 2014, Revised Selected Papers
PublisherSpringer Verlag
Pages88-98
Number of pages11
Volume8899
ISBN (Print)9783319134093
DOIs
StatePublished - 2014

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8899
ISSN (Print)03029743
ISSN (Electronic)16113349

Fingerprint

Feature Matching
Multi-model
Motion Estimation
Motion estimation
Scale Invariant Feature Transform
Motion
Model Verification
Empirical Analysis
Matching Algorithm
Labels
Cross-validation
Model
Visualization
Descriptors
Affine transformation
Cameras
Detectors
Correspondence
Camera
Detector

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Xiang, X., Mirota, D., Reiter, A., & Hager, G. (2014). Is multi-model feature matching better for endoscopic motion estimation? In Computer-Assisted and Robotic Endoscopy - 1st International Workshop, CARE 2014 held in Conjunction with MICCAI 2014, Revised Selected Papers (Vol. 8899, pp. 88-98). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8899). Springer Verlag. https://doi.org/10.1007/978-3-319-13410-9_9

Is multi-model feature matching better for endoscopic motion estimation? / Xiang, Xiang; Mirota, Daniel; Reiter, Austin; Hager, Gregory.

Computer-Assisted and Robotic Endoscopy - 1st International Workshop, CARE 2014 held in Conjunction with MICCAI 2014, Revised Selected Papers. Vol. 8899 Springer Verlag, 2014. p. 88-98 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8899).

Research output: Chapter in Book/Report/Conference proceedingChapter

Xiang, X, Mirota, D, Reiter, A & Hager, G 2014, Is multi-model feature matching better for endoscopic motion estimation? in Computer-Assisted and Robotic Endoscopy - 1st International Workshop, CARE 2014 held in Conjunction with MICCAI 2014, Revised Selected Papers. vol. 8899, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 8899, Springer Verlag, pp. 88-98. https://doi.org/10.1007/978-3-319-13410-9_9
Xiang X, Mirota D, Reiter A, Hager G. Is multi-model feature matching better for endoscopic motion estimation? In Computer-Assisted and Robotic Endoscopy - 1st International Workshop, CARE 2014 held in Conjunction with MICCAI 2014, Revised Selected Papers. Vol. 8899. Springer Verlag. 2014. p. 88-98. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-13410-9_9
Xiang, Xiang ; Mirota, Daniel ; Reiter, Austin ; Hager, Gregory. / Is multi-model feature matching better for endoscopic motion estimation?. Computer-Assisted and Robotic Endoscopy - 1st International Workshop, CARE 2014 held in Conjunction with MICCAI 2014, Revised Selected Papers. Vol. 8899 Springer Verlag, 2014. pp. 88-98 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inbook{346b5d8602154d9282fa124fb61208cf,
title = "Is multi-model feature matching better for endoscopic motion estimation?",
abstract = "Camera motion estimation is a standard yet critical step to endoscopic visualization. It is affected by the variation of locations and correspondences of features detected in 2D images. Feature detectors and descriptors vary, though one of the most widely used remains SIFT. Practitioners usually also adopt its feature matching strategy, which defines inliers as the feature pairs subjecting to a global affine transformation. However, for endoscopic videos, we are curious if it is more suitable to cluster features into multiple groups. We can still enforce the same transformation as in SIFT within each group. Such a multi-model idea has been recently examined in the Multi-Affine work, which outperforms Lowe’s SIFT in terms of re-projection error on minimally invasive endoscopic images with manually labelled ground-truth matches of SIFT features. Since their difference lies in matching, the accuracy gain of estimated motion is attributed to the holistic Multi-Affine feature matching algorithm. But, more concretely, the matching criterion and point searching can be the same as those built in SIFT. We argue that the real variation is only the motion model verification. We either enforce a single global motion model or employ a group of multiple local ones. In this paper, we investigate how sensitive the estimated motion is affected by the number of motion models assumed in feature matching. While the sensitivity can be analytically evaluated, we present an empirical analysis in a leaving-one-out cross validation setting without requiring labels of ground-truth matches. Then, the sensitivity is characterized by the variance of a sequence of motion estimates. We present a series of quantitative comparison such as accuracy and variance between Multi-Affine motion models and the global affine model.",
author = "Xiang Xiang and Daniel Mirota and Austin Reiter and Gregory Hager",
year = "2014",
doi = "10.1007/978-3-319-13410-9_9",
language = "English (US)",
isbn = "9783319134093",
volume = "8899",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "88--98",
booktitle = "Computer-Assisted and Robotic Endoscopy - 1st International Workshop, CARE 2014 held in Conjunction with MICCAI 2014, Revised Selected Papers",

}

TY - CHAP

T1 - Is multi-model feature matching better for endoscopic motion estimation?

AU - Xiang, Xiang

AU - Mirota, Daniel

AU - Reiter, Austin

AU - Hager, Gregory

PY - 2014

Y1 - 2014

N2 - Camera motion estimation is a standard yet critical step to endoscopic visualization. It is affected by the variation of locations and correspondences of features detected in 2D images. Feature detectors and descriptors vary, though one of the most widely used remains SIFT. Practitioners usually also adopt its feature matching strategy, which defines inliers as the feature pairs subjecting to a global affine transformation. However, for endoscopic videos, we are curious if it is more suitable to cluster features into multiple groups. We can still enforce the same transformation as in SIFT within each group. Such a multi-model idea has been recently examined in the Multi-Affine work, which outperforms Lowe’s SIFT in terms of re-projection error on minimally invasive endoscopic images with manually labelled ground-truth matches of SIFT features. Since their difference lies in matching, the accuracy gain of estimated motion is attributed to the holistic Multi-Affine feature matching algorithm. But, more concretely, the matching criterion and point searching can be the same as those built in SIFT. We argue that the real variation is only the motion model verification. We either enforce a single global motion model or employ a group of multiple local ones. In this paper, we investigate how sensitive the estimated motion is affected by the number of motion models assumed in feature matching. While the sensitivity can be analytically evaluated, we present an empirical analysis in a leaving-one-out cross validation setting without requiring labels of ground-truth matches. Then, the sensitivity is characterized by the variance of a sequence of motion estimates. We present a series of quantitative comparison such as accuracy and variance between Multi-Affine motion models and the global affine model.

AB - Camera motion estimation is a standard yet critical step to endoscopic visualization. It is affected by the variation of locations and correspondences of features detected in 2D images. Feature detectors and descriptors vary, though one of the most widely used remains SIFT. Practitioners usually also adopt its feature matching strategy, which defines inliers as the feature pairs subjecting to a global affine transformation. However, for endoscopic videos, we are curious if it is more suitable to cluster features into multiple groups. We can still enforce the same transformation as in SIFT within each group. Such a multi-model idea has been recently examined in the Multi-Affine work, which outperforms Lowe’s SIFT in terms of re-projection error on minimally invasive endoscopic images with manually labelled ground-truth matches of SIFT features. Since their difference lies in matching, the accuracy gain of estimated motion is attributed to the holistic Multi-Affine feature matching algorithm. But, more concretely, the matching criterion and point searching can be the same as those built in SIFT. We argue that the real variation is only the motion model verification. We either enforce a single global motion model or employ a group of multiple local ones. In this paper, we investigate how sensitive the estimated motion is affected by the number of motion models assumed in feature matching. While the sensitivity can be analytically evaluated, we present an empirical analysis in a leaving-one-out cross validation setting without requiring labels of ground-truth matches. Then, the sensitivity is characterized by the variance of a sequence of motion estimates. We present a series of quantitative comparison such as accuracy and variance between Multi-Affine motion models and the global affine model.

UR - http://www.scopus.com/inward/record.url?scp=84912135046&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84912135046&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-13410-9_9

DO - 10.1007/978-3-319-13410-9_9

M3 - Chapter

C2 - 26539567

AN - SCOPUS:84912135046

SN - 9783319134093

VL - 8899

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 88

EP - 98

BT - Computer-Assisted and Robotic Endoscopy - 1st International Workshop, CARE 2014 held in Conjunction with MICCAI 2014, Revised Selected Papers

PB - Springer Verlag

ER -