Learning photographic global tonal adjustment with a database of input/output image pairs

Bychkovsky, Vladimir; Paris, Sylvain; Chan, Eric K.; Durand, Frédo

doi:10.1109/cvpr.2011.5995413

Cited by 332 publications

(253 citation statements)

References 7 publications

Supporting

Mentioning

253

Contrasting

Order By: Relevance

“…Training data The number of images in our dataset is smaller than that in others (e.g., 1,300 in ours vs. 5,000 used in [9] from the MIT FiveK dataset [3]). However, our method gathers significantly more information per image than just input/output pairs.…”

Section: Discussionmentioning

confidence: 99%

“…These HDR tone mapping techniques have also been used to enhance local contrast in standard RGB images. Bychkovsky et al [3] presented a method for learning global tonal adjustments from a database of input-output image pairs. Different from the more general problem of color enhancement, tone mapping deals only with manipulating the luminance channel, mainly for the purpose of improving contrast.…”

Section: Related Workmentioning

confidence: 99%

“…The combination of an HSV histogram (6 bins) and RGB color moments is used as a descriptor. When there is no face in the image, we instead compute the histogram and moments over the entire image as done in [3].…”

Section: Color Featuresmentioning

confidence: 99%

“…For this, we use the 2D spatial Gaussian model of [3], fit to four intensity intervals but with four bins instead of ten.…”

Section: Color Featuresmentioning

confidence: 99%

“…Recent methods for enhancing the color of digital photographs have utilized image pairs for training, where each pair consists of an image before and after color adjustment by a human user [6,11,3,4]. With such a training set, a regression function is learned that maps the feature vector of a pre-adjusted image to its corresponding enhancements in color.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

A Learning-to-Rank Approach for Image Color Enhancement

Yan

Lin

Kang

et al. 2014

2014 IEEE Conference on Computer Vision and Pattern Recognition

View full text Add to dashboard Cite

We present a machine-learned ranking approach for automatically enhancing the color of a photograph. Unlike previous techniques that train on pairs of images before and after adjustment by a human user, our method takes into account the intermediate steps taken in the enhancement process, which provide detailed information on the person's color preferences. To make use of this data, we formulate the color enhancement task as a learning-to-rank problem in which ordered pairs of images are used for training, and then various color enhancements of a novel input image can be evaluated from their corresponding rank values. From the parallels between the decision tree structures we use for ranking and the decisions made by a human during the editing process, we posit that breaking a full enhancement sequence into individual steps can facilitate training. Our experiments show that this approach compares well to existing methods for automatic color enhancement.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Color Featuresmentioning

confidence: 99%

“…For this, we use the 2D spatial Gaussian model of [3], fit to four intensity intervals but with four bins instead of ten.…”

Section: Color Featuresmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

A Learning-to-Rank Approach for Image Color Enhancement

Yan

Lin

Kang

et al. 2014

2014 IEEE Conference on Computer Vision and Pattern Recognition

View full text Add to dashboard Cite

show abstract

Image naturalness model based on prototypical memory colors (sky, grass, and skin)

Tian

Luo

2022

Color Research & Application

View full text Add to dashboard Cite

Image naturalness is a potent attribute to enhance image quality, and it is related to people's perception of memory colors. Many previous experiments aimed to locate the memory colors on different media. Their findings indicate that the memory colors obtained vary between color patches and images and between experimental techniques. However, there were no satisfying models with appropriate memory color centers to scale the image naturalness. This study generated an image database including three prototypical memory colors images (sky, grass, and skin) and their local processed images. First, visual naturalness and preference data were collected from a psychophysical experiment. Then, the distribution of naturalness followed by memory color positions was analyzed. Based on the distribution pattern, the ellipsoid models for each prototypical memory color were established. The results from the present memory color centers were compared with those from the other studies. Results show that the present color centers always give better prediction accuracy in image naturalness models. Finally, an image preference model was developed based on the image naturalness and colorfulness.

show abstract

Flexible neural color compatibility model for efficient color extraction from image

Yan,

Xu,

Zhang

2023

Color Research & Application

View full text Add to dashboard Cite

Color choice is an essential aspect of many applications, including graphic design, web design and fashion design. The selection of colors can have a significant impact on the overall aesthetic and appeal of a design, as well as its effectiveness in conveying a particular message or mood. This paper introduces new and simple tools for choosing colors. First, we introduce a convolutional neural network that scores the quality of a set of five colors, called a color theme. Such a network can be used to rate the quality of a new color theme. Second, we propose a method to extract a variable‐size palette from an image. The size of the extracted palette can vary depending on the color richness of the image. Third, we demonstrate simple prototypes that apply the trained neural network and the palette extraction method to tasks in graphic design, such as improving existing themes. Our proposed network has the advantage of being significantly simpler than other state‐of‐the‐art methods with better performance.

show abstract

Learning photographic global tonal adjustment with a database of input/output image pairs

Cited by 332 publications

References 7 publications

A Learning-to-Rank Approach for Image Color Enhancement

A Learning-to-Rank Approach for Image Color Enhancement

Image naturalness model based on prototypical memory colors (sky, grass, and skin)

Flexible neural color compatibility model for efficient color extraction from image

Contact Info

Product

Resources

About