研究目的
Assessing the subjective preferences among different post-processing methods used to alleviate capture artifacts in low light consumer photos.
研究成果
Most people prefer sharpened images to denoised images even if sharpening amplifies the sensor noise. However, systematic differences in individual preferences can be observed, making automatic selection of optimal filter a challenging research topic.
研究不足
The study focused on consumer photos with typical capture artifacts in low light conditions, not aiming at the best possible technical quality. The selection of optimal filter is challenging due to systematic differences in individual preferences.
1：Experimental Design and Method Selection:
The study used a rank ordering method to compare subjective preferences between bilateral filter and sharpening methods on photos taken in low light conditions.
2：Sample Selection and Data Sources:
Photos were taken using both DSLR (Nikon D5100) and a smartphone (Huawei P9), mostly in low light conditions.
3：List of Experimental Equipment and Materials:
Nikon D5100 DSLR, Huawei P9 smartphone, test software developed for rank ordering.
4：Experimental Procedures and Operational Workflow:
Each source photo was cropped to 1920x960 pixel resolution and processed by applying bilateral filter for de-noising and unsharp mask for sharpening, both with two different settings.
5：Data Analysis Methods:
The results were analyzed in terms of Spearman Rank order Correlation Coefficient (SRCC) and Pearson Linear Correlation Coefficient (PLCC).
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
