Automatic red eye correction and its quality metric

Ilia V. Safonov; Michael N. Rychagov; Ki Min Kang; Sang Ho Kim

doi:10.1117/12.758603

Automatic red eye correction and its quality metric

Ilia V. Safonov, Michael N. Rychagov, Ki Min Kang, Sang Ho Kim

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

11 Scopus citations

Abstract

The red eye artifacts are troublesome defect of amateur photos. Correction of red eyes during printing without user intervention and making photos more pleasant for an observer are important tasks. The novel efficient technique of automatic correction of red eyes aimed for photo printers is proposed. This algorithm is independent from face orientation and capable to detect paired red eyes as well as single red eyes. The approach is based on application of 3D tables with typicalness levels for red eyes and human skin tones and directional edge detection filters for processing of redness image. Machine learning is applied for features selection. For classification of red eye regions a cascade of classifiers including Gentle AdaBoost committee from Classification and Regression Trees (CART) is applied. Retouching stage includes desaturation, darkening and blending with initial image. Several versions of approach implementation using trade-off between detection and correction quality, processing time, memory requirements are possible. The numeric quality criterion of automatic red eye correction is proposed. This quality metric is constructed by applying Analytic Hierarchy Process (AHP) to consumer opinions about correction outcomes. Proposed numeric metric helps to choose algorithm parameters via optimization procedure. Experimental results demonstrate high accuracy and efficiency of the proposed algorithm in comparison with existing solutions.

Original language	English
Title of host publication	Proceedings of SPIE-IS and T Electronic Imaging - Color Imaging XIII
Subtitle of host publication	Processing, Hardcopy, and Applications
DOIs	https://doi.org/10.1117/12.758603
State	Published - 2008
Externally published	Yes
Event	Color Imaging XIII: Processing, Hardcopy, and Applications - San Jose, CA, United States Duration: Jan 29 2008 → Jan 31 2008

Publication series

Name	Proceedings of SPIE - The International Society for Optical Engineering
Volume	6807
ISSN (Print)	0277-786X

Conference

Conference	Color Imaging XIII: Processing, Hardcopy, and Applications
Country/Territory	United States
City	San Jose, CA
Period	01/29/08 → 01/31/08

Keywords

Boosting
Classification
Directional edge detection filters
Quality metric
Red eye correction

Access to Document

10.1117/12.758603

Cite this

@inproceedings{638d1dbdd515412ea63ed73d93aba7db,

title = "Automatic red eye correction and its quality metric",

abstract = "The red eye artifacts are troublesome defect of amateur photos. Correction of red eyes during printing without user intervention and making photos more pleasant for an observer are important tasks. The novel efficient technique of automatic correction of red eyes aimed for photo printers is proposed. This algorithm is independent from face orientation and capable to detect paired red eyes as well as single red eyes. The approach is based on application of 3D tables with typicalness levels for red eyes and human skin tones and directional edge detection filters for processing of redness image. Machine learning is applied for features selection. For classification of red eye regions a cascade of classifiers including Gentle AdaBoost committee from Classification and Regression Trees (CART) is applied. Retouching stage includes desaturation, darkening and blending with initial image. Several versions of approach implementation using trade-off between detection and correction quality, processing time, memory requirements are possible. The numeric quality criterion of automatic red eye correction is proposed. This quality metric is constructed by applying Analytic Hierarchy Process (AHP) to consumer opinions about correction outcomes. Proposed numeric metric helps to choose algorithm parameters via optimization procedure. Experimental results demonstrate high accuracy and efficiency of the proposed algorithm in comparison with existing solutions.",

keywords = "Boosting, Classification, Directional edge detection filters, Quality metric, Red eye correction",

author = "Safonov, {Ilia V.} and Rychagov, {Michael N.} and Kang, {Ki Min} and Kim, {Sang Ho}",

year = "2008",

doi = "10.1117/12.758603",

language = "English",

isbn = "9780819469793",

series = "Proceedings of SPIE - The International Society for Optical Engineering",

booktitle = "Proceedings of SPIE-IS and T Electronic Imaging - Color Imaging XIII",

note = "Color Imaging XIII: Processing, Hardcopy, and Applications ; Conference date: 29-01-2008 Through 31-01-2008",

}

Safonov, IV, Rychagov, MN, Kang, KM & Kim, SH 2008, Automatic red eye correction and its quality metric. in Proceedings of SPIE-IS and T Electronic Imaging - Color Imaging XIII: Processing, Hardcopy, and Applications., 68070W, Proceedings of SPIE - The International Society for Optical Engineering, vol. 6807, Color Imaging XIII: Processing, Hardcopy, and Applications, San Jose, CA, United States, 01/29/08. https://doi.org/10.1117/12.758603

Automatic red eye correction and its quality metric. / Safonov, Ilia V.; Rychagov, Michael N.; Kang, Ki Min et al.
Proceedings of SPIE-IS and T Electronic Imaging - Color Imaging XIII: Processing, Hardcopy, and Applications. 2008. 68070W (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 6807).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Automatic red eye correction and its quality metric

AU - Safonov, Ilia V.

AU - Rychagov, Michael N.

AU - Kang, Ki Min

AU - Kim, Sang Ho

PY - 2008

Y1 - 2008

N2 - The red eye artifacts are troublesome defect of amateur photos. Correction of red eyes during printing without user intervention and making photos more pleasant for an observer are important tasks. The novel efficient technique of automatic correction of red eyes aimed for photo printers is proposed. This algorithm is independent from face orientation and capable to detect paired red eyes as well as single red eyes. The approach is based on application of 3D tables with typicalness levels for red eyes and human skin tones and directional edge detection filters for processing of redness image. Machine learning is applied for features selection. For classification of red eye regions a cascade of classifiers including Gentle AdaBoost committee from Classification and Regression Trees (CART) is applied. Retouching stage includes desaturation, darkening and blending with initial image. Several versions of approach implementation using trade-off between detection and correction quality, processing time, memory requirements are possible. The numeric quality criterion of automatic red eye correction is proposed. This quality metric is constructed by applying Analytic Hierarchy Process (AHP) to consumer opinions about correction outcomes. Proposed numeric metric helps to choose algorithm parameters via optimization procedure. Experimental results demonstrate high accuracy and efficiency of the proposed algorithm in comparison with existing solutions.

AB - The red eye artifacts are troublesome defect of amateur photos. Correction of red eyes during printing without user intervention and making photos more pleasant for an observer are important tasks. The novel efficient technique of automatic correction of red eyes aimed for photo printers is proposed. This algorithm is independent from face orientation and capable to detect paired red eyes as well as single red eyes. The approach is based on application of 3D tables with typicalness levels for red eyes and human skin tones and directional edge detection filters for processing of redness image. Machine learning is applied for features selection. For classification of red eye regions a cascade of classifiers including Gentle AdaBoost committee from Classification and Regression Trees (CART) is applied. Retouching stage includes desaturation, darkening and blending with initial image. Several versions of approach implementation using trade-off between detection and correction quality, processing time, memory requirements are possible. The numeric quality criterion of automatic red eye correction is proposed. This quality metric is constructed by applying Analytic Hierarchy Process (AHP) to consumer opinions about correction outcomes. Proposed numeric metric helps to choose algorithm parameters via optimization procedure. Experimental results demonstrate high accuracy and efficiency of the proposed algorithm in comparison with existing solutions.

KW - Boosting

KW - Classification

KW - Directional edge detection filters

KW - Quality metric

KW - Red eye correction

UR - http://www.scopus.com/inward/record.url?scp=45549104595&partnerID=8YFLogxK

U2 - 10.1117/12.758603

DO - 10.1117/12.758603

M3 - Conference contribution

AN - SCOPUS:45549104595

SN - 9780819469793

T3 - Proceedings of SPIE - The International Society for Optical Engineering

BT - Proceedings of SPIE-IS and T Electronic Imaging - Color Imaging XIII

T2 - Color Imaging XIII: Processing, Hardcopy, and Applications

Y2 - 29 January 2008 through 31 January 2008

ER -

Automatic red eye correction and its quality metric

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this