Using words as lexical basis functions for automatically indexing face images in a manner that correlates with human perception of similarity

Mariano Phielipp, John A. Black, Sethuraman Panchanathan

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

To facilitate collaboration between computers and people, computers should be able to perceive the world in a manner that correlates well with human perception. A good example of this is face image retrieval. Mathematically-based face indexing methods that are not based primarily on how humans perceive faces can produce retrievals mat are disappointing to human users. This raises the question "Can human faces be automatically indexed in a manner that correlates well with human perception of similarity?" Humans use words to describe faces - words such as braided, graybearded, bearded, bespectacled, bald, blondish, blond, freckled, blue eyed, mustached, pale, Caucasian, brown eyed, dark skinned, or black eyed. Such words represent dimensions that span a shared concept space for faces. Therefore they might provide a useful guide to indexing faces in an intuitive manner. This paper describes research that uses descriptive words such as these to index faces. Each word guides the design of one feature detector that produces a scalar coefficient, and those coefficients collectively define a feature vector for each face. Given these feature vectors, it is possible to compute a similarity measure between pairs of faces, and to compare that computed similarity to the similarity, as perceived by humans.

Original languageEnglish (US)
Title of host publicationHuman Vision and Electronic Imaging XI - Proceedings of SPIE-IS and T Electronic Imaging
Volume6057
DOIs
StatePublished - 2006
EventHuman Vision and Electronic Imaging XI - San Jose, CA, United States
Duration: Jan 16 2006Jan 18 2006

Other

OtherHuman Vision and Electronic Imaging XI
CountryUnited States
CitySan Jose, CA
Period1/16/061/18/06

Fingerprint

Image retrieval
retrieval
Detectors
coefficients
scalars
detectors

Keywords

  • Face indexing
  • Face recognition
  • Facial similarity
  • Feature detectors
  • Human visual perception
  • Human-computer collaboration
  • Lexical basis functions

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Condensed Matter Physics

Cite this

Phielipp, M., Black, J. A., & Panchanathan, S. (2006). Using words as lexical basis functions for automatically indexing face images in a manner that correlates with human perception of similarity. In Human Vision and Electronic Imaging XI - Proceedings of SPIE-IS and T Electronic Imaging (Vol. 6057). [60571C] https://doi.org/10.1117/12.643692

Using words as lexical basis functions for automatically indexing face images in a manner that correlates with human perception of similarity. / Phielipp, Mariano; Black, John A.; Panchanathan, Sethuraman.

Human Vision and Electronic Imaging XI - Proceedings of SPIE-IS and T Electronic Imaging. Vol. 6057 2006. 60571C.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Phielipp, M, Black, JA & Panchanathan, S 2006, Using words as lexical basis functions for automatically indexing face images in a manner that correlates with human perception of similarity. in Human Vision and Electronic Imaging XI - Proceedings of SPIE-IS and T Electronic Imaging. vol. 6057, 60571C, Human Vision and Electronic Imaging XI, San Jose, CA, United States, 1/16/06. https://doi.org/10.1117/12.643692
Phielipp M, Black JA, Panchanathan S. Using words as lexical basis functions for automatically indexing face images in a manner that correlates with human perception of similarity. In Human Vision and Electronic Imaging XI - Proceedings of SPIE-IS and T Electronic Imaging. Vol. 6057. 2006. 60571C https://doi.org/10.1117/12.643692
Phielipp, Mariano ; Black, John A. ; Panchanathan, Sethuraman. / Using words as lexical basis functions for automatically indexing face images in a manner that correlates with human perception of similarity. Human Vision and Electronic Imaging XI - Proceedings of SPIE-IS and T Electronic Imaging. Vol. 6057 2006.
@inproceedings{427a996e356f4831bc7635fe58d48714,
title = "Using words as lexical basis functions for automatically indexing face images in a manner that correlates with human perception of similarity",
abstract = "To facilitate collaboration between computers and people, computers should be able to perceive the world in a manner that correlates well with human perception. A good example of this is face image retrieval. Mathematically-based face indexing methods that are not based primarily on how humans perceive faces can produce retrievals mat are disappointing to human users. This raises the question {"}Can human faces be automatically indexed in a manner that correlates well with human perception of similarity?{"} Humans use words to describe faces - words such as braided, graybearded, bearded, bespectacled, bald, blondish, blond, freckled, blue eyed, mustached, pale, Caucasian, brown eyed, dark skinned, or black eyed. Such words represent dimensions that span a shared concept space for faces. Therefore they might provide a useful guide to indexing faces in an intuitive manner. This paper describes research that uses descriptive words such as these to index faces. Each word guides the design of one feature detector that produces a scalar coefficient, and those coefficients collectively define a feature vector for each face. Given these feature vectors, it is possible to compute a similarity measure between pairs of faces, and to compare that computed similarity to the similarity, as perceived by humans.",
keywords = "Face indexing, Face recognition, Facial similarity, Feature detectors, Human visual perception, Human-computer collaboration, Lexical basis functions",
author = "Mariano Phielipp and Black, {John A.} and Sethuraman Panchanathan",
year = "2006",
doi = "10.1117/12.643692",
language = "English (US)",
isbn = "0819460974",
volume = "6057",
booktitle = "Human Vision and Electronic Imaging XI - Proceedings of SPIE-IS and T Electronic Imaging",

}

TY - GEN

T1 - Using words as lexical basis functions for automatically indexing face images in a manner that correlates with human perception of similarity

AU - Phielipp, Mariano

AU - Black, John A.

AU - Panchanathan, Sethuraman

PY - 2006

Y1 - 2006

N2 - To facilitate collaboration between computers and people, computers should be able to perceive the world in a manner that correlates well with human perception. A good example of this is face image retrieval. Mathematically-based face indexing methods that are not based primarily on how humans perceive faces can produce retrievals mat are disappointing to human users. This raises the question "Can human faces be automatically indexed in a manner that correlates well with human perception of similarity?" Humans use words to describe faces - words such as braided, graybearded, bearded, bespectacled, bald, blondish, blond, freckled, blue eyed, mustached, pale, Caucasian, brown eyed, dark skinned, or black eyed. Such words represent dimensions that span a shared concept space for faces. Therefore they might provide a useful guide to indexing faces in an intuitive manner. This paper describes research that uses descriptive words such as these to index faces. Each word guides the design of one feature detector that produces a scalar coefficient, and those coefficients collectively define a feature vector for each face. Given these feature vectors, it is possible to compute a similarity measure between pairs of faces, and to compare that computed similarity to the similarity, as perceived by humans.

AB - To facilitate collaboration between computers and people, computers should be able to perceive the world in a manner that correlates well with human perception. A good example of this is face image retrieval. Mathematically-based face indexing methods that are not based primarily on how humans perceive faces can produce retrievals mat are disappointing to human users. This raises the question "Can human faces be automatically indexed in a manner that correlates well with human perception of similarity?" Humans use words to describe faces - words such as braided, graybearded, bearded, bespectacled, bald, blondish, blond, freckled, blue eyed, mustached, pale, Caucasian, brown eyed, dark skinned, or black eyed. Such words represent dimensions that span a shared concept space for faces. Therefore they might provide a useful guide to indexing faces in an intuitive manner. This paper describes research that uses descriptive words such as these to index faces. Each word guides the design of one feature detector that produces a scalar coefficient, and those coefficients collectively define a feature vector for each face. Given these feature vectors, it is possible to compute a similarity measure between pairs of faces, and to compare that computed similarity to the similarity, as perceived by humans.

KW - Face indexing

KW - Face recognition

KW - Facial similarity

KW - Feature detectors

KW - Human visual perception

KW - Human-computer collaboration

KW - Lexical basis functions

UR - http://www.scopus.com/inward/record.url?scp=33645523474&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33645523474&partnerID=8YFLogxK

U2 - 10.1117/12.643692

DO - 10.1117/12.643692

M3 - Conference contribution

AN - SCOPUS:33645523474

SN - 0819460974

SN - 9780819460974

VL - 6057

BT - Human Vision and Electronic Imaging XI - Proceedings of SPIE-IS and T Electronic Imaging

ER -