English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Portrait photo retouching is a photo retouching task that emphasizeshuman-region priority and group-level consistency. The lookup table-basedmethod achieves promising retouching performance by learning image-adaptiveweights to combine 3-dimensional lookup tables (3D LUTs) and conductingpixel-to-pixel color transformation. However, this paradigm ignores the localcontext cues and applies the same transformation to portrait pixels andbackground pixels when they exhibit the same raw RGB values. In contrast, anexpert usually conducts different operations to adjust the color temperaturesand tones of portrait regions and background regions. This inspires us to modellocal context cues to improve the retouching quality explicitly. Firstly, weconsider an image patch and predict pixel-adaptive lookup table weights toprecisely retouch the center pixel. Secondly, as neighboring pixels exhibitdifferent affinities to the center pixel, we estimate a local attention mask tomodulate the influence of neighboring pixels. Thirdly, the quality of the localattention mask can be further improved by applying supervision, which is basedon the affinity map calculated by the groundtruth portrait mask. As forgroup-level consistency, we propose to directly constrain the variance of meancolor components in the Lab space. Extensive experiments on PPR10K datasetverify the effectiveness of our method, e.g. on high-resolution photos, thePSNR metric receives over 0.5 gains while the group-level consistency metricobtains at least 2.1 decreases.

Learning Pixel-Adaptive Weights for Portrait Photo Retouching