Computer Laboratory

Projects

CameraVDP: Perceptual Display Assessment with Uncertainty Estimation via Camera and Visual Difference Prediction

Presented at SIGGRAPH Asia 2025

Yancheng Cai(1), Robert Wanat(2), and Rafał K. Mantiuk(1).

(1) University of Cambridge (2) LG Electronics North America


Contact Email: yc613 [AT] cam.ac.uk
Contact WeChat: cyc13700232963


Abstract

Accurate measurement of images produced by electronic displays is critical for the evaluation of both traditional and computational displays. Traditional display measurement methods based on sparse radiometric sampling and fitting a model are inadequate for capturing spatially varying display artifacts, as they fail to capture high-frequency and pixel-level distortions. While cameras offer sufficient spatial resolution, they introduce optical, sampling, and photometric distortions. Furthermore, the physical measurement must be combined with a model of a visual system to assess whether the distortions are going to be visible. To enable perceptual assessment of displays, we propose a combination of a camera-based reconstruction pipeline with a visual difference predictor, which account for both the inaccuracy of camera measurements and visual difference prediction. The reconstruction pipeline combines HDR image stacking, MTF inversion, vignetting correction, geometric undistortion, homography transformation, and color correction, enabling cameras to function as precise display measurement instruments. By incorporating a Visual Difference Predictor (VDP), our system models the visibility of various stimuli under different viewing conditions for the human visual system. We validate the proposed CameraVDP framework through three applications: defective pixel detection, color fringing awareness, and display non-uniformity evaluation. Our uncertainty analysis framework enables the estimation of the theoretical upper bound for defect pixel detection performance and provides confidence intervals for VDP quality scores.

Camera Measurement Pipeline
The complete camera measurement pipeline is shown from left to right, with each column representing a key step and the bottom row illustrating transformation examples. Column 1: HDR acquisition via merging multi-exposure images (4 exposure levels). Column 2: MTF inversion to correct lens aberrations. Column 3: Vignetting correction, exemplified by the vignetting map of the Sony α7R III with FE 1.8/35mm lens. Column 4: Intrinsic and distortion parameter estimation using full-screen checkerboard patterns for geometric undistortion (crop empty edges caused by undistortion). Column 5: Homography estimation of extrinsic parameters using full-screen OpenCV ArUco markers. The display pixel oversampling factor o denotes the distance, in image pixels, between the centers of original display pixels (red squares) in the oversampled image after the homography transformation. Examples with o = 2, 4 are shown, where red squares indicate display pixel centers. Column 6: Color correction by mapping camera RGB to measured XYZ.

Materials

  • Paper:
    CameraVDP: Perceptual Display Assessment with Uncertainty Estimation via Camera and Visual Difference Prediction
    Yancheng Cai, Robert Wanat, Rafał K. Mantiuk.
    In SIGGRAPH Asia 2025 Papers
    [arxiv] [paper PDF]
  • Code [Github]

Related projects

  • ColorVideoVDP - ColorVideoVDP: A visual difference predictor for image, video and display distortions
  • castleCSF - A Contrast Sensitivity Function of Color, Area, Spatio-Temporal frequency, Luminance and Eccentricity - models contrast sensitivity in ColorVideoVDP
  • elaTCSF - A Temporal Contrast Sensitivity Function for Flicker Detection and Modeling Variable Refresh Rate Flicker