研究目的
To present new measurement metrics, test scenes, and a new protocol to assess the quality of digital cameras using HDR technologies.
研究成果
The paper presents a novel laboratory setup and improved image quality measures for evaluating HDR technologies in digital cameras. The measures obtained are independent of the scene content and were validated through benchmarking and subjective evaluations. The study highlights the evolution of HDR technology over time and identifies key findings from the proposed HDR measures.
研究不足
The entropy-based measure for contrast preservation does not fully capture human preference for slightly saturated images. The study also notes that the repeatability of the illumination and print properties of the back-lit HDR setup is not as good as that of the ColorChecker.
1：Experimental Design and Method Selection:
The study uses a laboratory setup with two programmable light panels and printed transparent charts to create reproducible high dynamic range scenes. The setup allows for the measurement of local contrast, texture, color consistency, and noise in HDR images.
2：Sample Selection and Data Sources:
Four leading smartphone cameras launched between 2014 and 2016 were benchmarked. The study also includes subjective evaluations from fifteen subjects on six natural HDR scenes.
3：List of Experimental Equipment and Materials:
Two diffuse and adjustable light sources (Kino Flo LED 201 DMX devices), transparent prints containing a grayscale, a color, and a texture chart.
4：Experimental Procedures and Operational Workflow:
The two DMX devices are programmed to vary the luminous emittance, creating scenes with increasing dynamic range. For each dynamic range setting, a photograph is acquired with the HDR setting and automatic exposure.
5：Data Analysis Methods:
The study uses entropy for contrast preservation, acutance for texture preservation, color difference in the CIE L*a*b* color space for color consistency, and visual noise for noise analysis.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
