Effect of camera resolution and bandwidth on facial affect recognition

Mario Cruz, Robyn Flaum Cruz, Elizabeth A. Krupinski, Ana Maria Lopez, Richard M. McNeeley, Ronald S. Weinstein

Research output: Contribution to journalArticle

11 Scopus citations

Abstract

This preliminary study explored the effect of camera resolution and bandwidth on facial affect recognition, an important process and clinical variable in mental health service delivery. Sixty medical students and mental health-care professionals were recruited and randomized to four different combinations of commonly used teleconferencing camera resolutions and bandwidths: (1) one chip charged coupling device (CCD) camera, commonly used for VHS-grade taping and in teleconferencing systems costing less than $4,000 with a resolution of 280 lines, and 128 kilobytes per second bandwidth (kbps); (2) VHS and 768 kbps; (3) three-chip CCD camera, commonly used for Betacam (Beta) grade taping and in teleconferencing systems costing more than $4,000 with a resolution of 480 lines, and 128 kbps; and (4) Betacam and 768 kbps. The subjects were asked to identify four facial affects dynamically presented on videotape by an actor and actress presented via a video monitor at 30 frames per second. Two-way analysis of variance (ANOVA) revealed a significant interaction effect for camera resolution and bandwidth (p = 0.02) and a significant main effect for camera resolution (p = 0.006), but no main effect for bandwidth was detected. Post hoc testing of interaction means, using the Tukey Honestly Significant Difference (HSD) test and the critical difference (CD) at the 0.05 alpha level = 1.71, revealed subjects in the VHS/768 kbps (M = 7.133) and VHS/128 kbps (M = 6.533) were significantly better at recognizing the displayed facial affects than those in the Betacam/768 kbps (M = 4.733) or Betacam/128 kbps (M = 6.333) conditions. Camera resolution and bandwidth combinations differ in their capacity to influence facial affect recognition. For service providers, this study's results support the use of VHS cameras with either 768 kbps or 128 kbps bandwidths for facial affect recognition compared to Betacam cameras. The authors argue that the results of this study are a consequence of the VHS camera resolution/bandwidth combinations' ability to improve signal detection (i.e., facial affect recognition) by subjects in comparison to Betacam camera resolution/bandwidth combinations.

Original languageEnglish (US)
Pages (from-to)392-402
Number of pages11
JournalTelemedicine Journal and e-Health
Volume10
Issue number3
DOIs
StatePublished - Sep 1 2004

ASJC Scopus subject areas

  • Health Informatics
  • Health Information Management

Fingerprint Dive into the research topics of 'Effect of camera resolution and bandwidth on facial affect recognition'. Together they form a unique fingerprint.

  • Cite this