6533b7d7fe1ef96bd12676d6
RESEARCH PRODUCT
Camera-based measurement of relative image contrast in projection displays
Ping Zhao Marius Pedersen Jon Yngve Hardeberg Jean-baptiste Thomassubject
[ INFO.INFO-TS ] Computer Science [cs]/Signal and Image Processingoptical projectors[INFO.INFO-TS] Computer Science [cs]/Signal and Image Processingprojection luminanceComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION[ SPI.SIGNAL ] Engineering Sciences [physics]/Signal and Image processingmeasured contrastradiometersImage color analysismetrics[INFO.INFO-TS]Computer Science [cs]/Signal and Image Processingstatistical analysispicture acquisitioncamerasgray scaleBrightnessoptical variables measurementdigital cameracamera-based measurementRadiometry[SPI.SIGNAL] Engineering Sciences [physics]/Signal and Image processingdisplay devicesDigital camerasuncalibrated digital cameraglobal contrast metricsprojector brightness settingscheckerboardcolor complex test imagesrelative image contrastviewing conditionsradiometerimage processingCorrelationPearson correlationhigh-end radiometerprojection displayprojector contrast settings[SPI.SIGNAL]Engineering Sciences [physics]/Signal and Image processingprojection displayscontrast surface plotsstatistic based metricsdescription
International audience; This research investigated the measured contrast of projection displays based on pictures taken by un-calibrated digital cameras under typical viewing conditions. A high-end radiometer was employed as a reference to the physical response of projection luminance. Checkerboard, gray scale and color complex test images with a range of the projector's brightness and contrast settings were projected. Two local and two global contrast metrics were evaluated on the acquired pictures. We used contrast surface plots and Pearson correlation to investigate the measured contrast versus the projector's brightness and contrast settings. The results suggested, as expected, the projector contrast has a more significant impact on measured contrast than projector brightness, but the measured contrast based on either camera or radiometer has a nonlinear relationship with projector settings. The results also suggested that simple statistics based metrics might produce a higher Pearson correlation value with both projector contrast and projector brightness than more complex contrast metrics. Our results demonstrated that the rank order of un-calibrated camera based measured contrast and radiometer based measured contrast is preserved for large steps of projector setting differences.
| year | journal | country | edition | language |
|---|---|---|---|---|
| 2013-06-10 |