Unsupervised calibration of RGB-NIR capture pairs utilizing dense multimodal image correspondences

Filipe Gama; Mihail Georgiev; Atanas Gotchev

doi:10.23919/EUSIPCO.2018.8553454

Unsupervised calibration of RGB-NIR capture pairs utilizing dense multimodal image correspondences

Filipe Gama
, Mihail Georgiev
, Atanas Gotchev

Tutkimustuotos: Konferenssiartikkeli › Tieteellinen › vertaisarvioitu

Abstrakti

In this paper, we propose an unsupervised calibration framework aimed at calibrating RGB plus Near-InfraRed (NIR) capture setups. We favour dense feature matching for the case of multimodal data and utilize the Scale-Invariant Feature Transform (SIFT) flow, previously developed for matching same-category image objects. We develop an optimization procedure that minimizes the global disparity field between the two multimodal images in order to adapt SIFT flow for our calibration needs. The proposed optimization substantially increases the number of inliers and yields more robust and unambiguous calibration results.

Alkuperäiskieli	Englanti
Otsikko	2018 26th European Signal Processing Conference (EUSIPCO)
Kustantaja	IEEE
Sivut	2145-2149
Sivumäärä	5
ISBN (elektroninen)	978-9-0827-9701-5
ISBN (painettu)	978-1-5386-3736-4
DOI - pysyväislinkit	https://doi.org/10.23919/EUSIPCO.2018.8553454
Tila	Julkaistu - syysk. 2018
OKM-julkaisutyyppi	A4 Artikkeli konferenssijulkaisussa
Tapahtuma	EUROPEAN SIGNAL PROCESSING CONFERENCE - Kesto: 1 tammik. 1900 → …

Julkaisusarja

Nimi
ISSN (elektroninen)	2076-1465

Conference

Conference	EUROPEAN SIGNAL PROCESSING CONFERENCE
Ajanjakso	1/01/00 → …

Julkaisufoorumi-taso

Jufo-taso 1

Pääsy asiakirjaan

10.23919/EUSIPCO.2018.8553454

26th European Signal Processing Conference
Da Graca Gama, F. (Osallistuja)
3 syysk. 2018 → 7 syysk. 2018
Aktiviteetti: Konferenssiin, työryhmään, paneeliin, sessioon tai tutoriaaliin osallistuminen

Siteeraa tätä

@inproceedings{033287d0006142c9aec40e054f8bf9d4,

title = "Unsupervised calibration of RGB-NIR capture pairs utilizing dense multimodal image correspondences",

abstract = "In this paper, we propose an unsupervised calibration framework aimed at calibrating RGB plus Near-InfraRed (NIR) capture setups. We favour dense feature matching for the case of multimodal data and utilize the Scale-Invariant Feature Transform (SIFT) flow, previously developed for matching same-category image objects. We develop an optimization procedure that minimizes the global disparity field between the two multimodal images in order to adapt SIFT flow for our calibration needs. The proposed optimization substantially increases the number of inliers and yields more robust and unambiguous calibration results.",

keywords = "calibration, feature extraction, image colour analysis, image matching, optimisation, transforms, SIFT flow, unambiguous calibration results, optimization procedure, same-category image objects, multimodal data, dense feature matching, unsupervised calibration framework, dense multimodal image correspondences, RGB-NIR capture pairs, global disparity field minimization, RGB-plus-near-infrared capture setups, scale-invariant feature transform flow, Cameras, Calibration, Feature extraction, Matched filters, Optimization, Sensors, Genetic algorithms, NIR, multimodal stereo, features matching",

author = "Filipe Gama and Mihail Georgiev and Atanas Gotchev",

year = "2018",

month = sep,

doi = "10.23919/EUSIPCO.2018.8553454",

language = "English",

isbn = "978-1-5386-3736-4",

publisher = "IEEE",

pages = "2145--2149",

booktitle = "2018 26th European Signal Processing Conference (EUSIPCO)",

address = "United States",

note = "European Signal Processing Conference ; Conference date: 01-01-1900",

}

TY - GEN

T1 - Unsupervised calibration of RGB-NIR capture pairs utilizing dense multimodal image correspondences

AU - Gama, Filipe

AU - Georgiev, Mihail

AU - Gotchev, Atanas

PY - 2018/9

Y1 - 2018/9

N2 - In this paper, we propose an unsupervised calibration framework aimed at calibrating RGB plus Near-InfraRed (NIR) capture setups. We favour dense feature matching for the case of multimodal data and utilize the Scale-Invariant Feature Transform (SIFT) flow, previously developed for matching same-category image objects. We develop an optimization procedure that minimizes the global disparity field between the two multimodal images in order to adapt SIFT flow for our calibration needs. The proposed optimization substantially increases the number of inliers and yields more robust and unambiguous calibration results.

AB - In this paper, we propose an unsupervised calibration framework aimed at calibrating RGB plus Near-InfraRed (NIR) capture setups. We favour dense feature matching for the case of multimodal data and utilize the Scale-Invariant Feature Transform (SIFT) flow, previously developed for matching same-category image objects. We develop an optimization procedure that minimizes the global disparity field between the two multimodal images in order to adapt SIFT flow for our calibration needs. The proposed optimization substantially increases the number of inliers and yields more robust and unambiguous calibration results.

KW - calibration

KW - feature extraction

KW - image colour analysis

KW - image matching

KW - optimisation

KW - transforms

KW - SIFT flow

KW - unambiguous calibration results

KW - optimization procedure

KW - same-category image objects

KW - multimodal data

KW - dense feature matching

KW - unsupervised calibration framework

KW - dense multimodal image correspondences

KW - RGB-NIR capture pairs

KW - global disparity field minimization

KW - RGB-plus-near-infrared capture setups

KW - scale-invariant feature transform flow

KW - Cameras

KW - Calibration

KW - Feature extraction

KW - Matched filters

KW - Optimization

KW - Sensors

KW - Genetic algorithms

KW - NIR

KW - multimodal stereo

KW - features matching

U2 - 10.23919/EUSIPCO.2018.8553454

DO - 10.23919/EUSIPCO.2018.8553454

M3 - Conference contribution

SN - 978-1-5386-3736-4

SP - 2145

EP - 2149

BT - 2018 26th European Signal Processing Conference (EUSIPCO)

PB - IEEE

T2 - European Signal Processing Conference

Y2 - 1 January 1900

ER -

Unsupervised calibration of RGB-NIR capture pairs utilizing dense multimodal image correspondences

Abstrakti

Julkaisusarja

Conference

Julkaisufoorumi-taso

Pääsy asiakirjaan

Sormenjälki

Aktiviteetit

26th European Signal Processing Conference

Siteeraa tätä