Apparent Color Dataset: How Apparent Color Differs from the Color Extracted from Photos

Kubota, Yuki; Yoshida, Shigeo; İnami, Masahiko

doi:10.1109/smc53654.2022.9945597

Cited by 1 publication

(5 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Firstly, color differences might vary between categories. Analyzing the apparent color dataset (Kubota et al, 2022), the median color differences between the apparent color and pixel color for Categories 0 and 2 were 6.89 and 6.03. Conversely, for Categories 3 and 4, which exhibited superior prediction accuracy, these values increased to 12.7 and 11.8.…”

Section: Discussionmentioning

confidence: 99%

“…However, these studies do not consider human color perception that are biased in both the pixel-wise color and the perfect whitebalanced colors. This research aims to construct the regression model that predicts the users' apparent colors in natural photographs based on a psychophysical dataset named an apparent color dataset (Kubota et al, 2022). Note that the main focus of this paper is prediction model construction and evaluation, which differs from the purpose of the previous paper, which focused on dataset construction and its analysis.…”

Section: Color Picker As An Interactive Interfacementioning

confidence: 99%

“…The apparent color dataset (Kubota et al, 2022) was utilized for training the proposed model. The dataset consists of 1,100 photographs, each annotated with the apparent color using the adjustment method with a color palette.…”

Section: Performance Evaluationmentioning

confidence: 99%

“…Dataset with pronounced color differences in the apparent color dataset (Kubota et al, 2022) may have been caused by participants' failure to adjust their apparent color properly. Owing to the potential risk of incorporating such outlier data on prediction accuracy, the authors examined Point 3.…”

Section: What Is the Appropriate Range Of Training Data? (Point 3)mentioning

confidence: 99%

“…The candidate stimuli were first collected as 10,000 imageposition pairs from the DTD dataset and 20,000 image-position pairs from the Places dataset, following the previous paper constructing the apparent color dataset (Kubota et al, 2022). The pairs of image and response position were selected so that the color difference ΔE 00 < 5.0 in the range of 5 px surroundings of the position to be responded to.…”

Section: Stimulimentioning

confidence: 99%

See 4 more Smart Citations

Apparent color picker: color prediction model to extract apparent color in photos

Kubota

Yoshida

İnami

2023

Front. Signal Process.

Self Cite

View full text Add to dashboard Cite

A color extraction interface reflecting human color perception helps pick colors from natural images as users see. Apparent color in photos differs from pixel color due to complex factors, including color constancy and adjacent color. However, methodologies for estimating the apparent color in photos have yet to be proposed. In this paper, the authors investigate suitable model structures and features for constructing an apparent color picker, which extracts the apparent color from natural photos. Regression models were constructed based on the psychophysical dataset for given images to predict the apparent color from image features. The linear regression model incorporates features that reflect multi-scale adjacent colors. The evaluation experiments confirm that the estimated color was closer to the apparent color than the pixel color for an average of 70%–80% of the images. However, the accuracy decreased for several conditions, including low and high saturation at low luminance. The authors believe that the proposed methodology could be applied to develop user interfaces to compensate for the discrepancy between human perception and computer predictions.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Color Picker As An Interactive Interfacementioning

confidence: 99%

Section: Performance Evaluationmentioning

confidence: 99%

Section: What Is the Appropriate Range Of Training Data? (Point 3)mentioning

confidence: 99%

Section: Stimulimentioning

confidence: 99%

See 3 more Smart Citations