Independent control of tonescale regions (e.g., shadows, highlights) is essential for painters, photographers and cinematographers to bring 2D images to life. In image manipulation software this is most directly addressed by color grading modules, which use intensity thresholds to segment distinct illumination regions for local manipulation. In this work we develop an open source color grading tool and use it to annotate a large dataset of video frames with tonescale region thresholds. Using these thresholds we conduct modeling experiments with strategies based on both practitioners' conventional wisdom and machine learning. Results show that K-nearest neighbors is an effective prediction strategy, outperforming state-of-the-art end-to-end methods for image enhancement. This outcome demonstrates the benefit of focusing on a compact set of core parameters when modeling creative stylization processes. Our adaptive color grading interface and data are available at https://github.com/SamsungLabs/adaptive-color-grading.
Figures & tables
Figure 1: A relatively “flat” image appears more voluminous when the contrast between its tonescale regions is increased 1,2 . This can be accomplished through tonemapping and chromatic shifts 2,3 . However, a static 3D color lookup table tuned to achieve this effect for one image may not work for another 4,5 . We present a method for predicting tonescale region thresholds such that these color effects can be adapted to new images 4,6 .
Figure 2: Color grading interface featuring chroma offset, TRT controls and grayscale visualization.
Figure 3: Color grading architecture. Regional LUTs are applied in a nested fashion to an identity LUT which is then applied to return the output image.
Figure 4: The HDR videographic survey dataset arranged according to scene type.
Figure 5: Histogram of tonescale region settings.
Location subset
Count
Scene type subsets
Toronto
90
Urban
Utah
458
Sunset, Night
Alaska
378
Sunset, Night
Seattle
286
Sunset, Night, Urban, Indoor
London
146
Sunset, Night, Urban, Indoor
NYC
206
Sunset, Urban, Indoor
Table 1: Breakdown of HVS locations and scene types which differ from the majority day, outdoor, nature scenes.
Figure 6: 1-D Optimization landscapes for Fixed and Percentile based strategies.
Table 2: Quantitative results of modeling approaches versus end-to-end deep learning approaches. In these experiments models are tested on a subset and trained on the remainder or vice-versa.
Figure 9: Top row: 90th percentile ΔE00 averaged across the test set (x-axis). Bottom row: Correlation plots with ground truth darkest (purple) dark (blue) light (green) lightest (red) TRTs plotted against predicted ones by KNN, MLP, Fixed and Percentile strategies.
Figure 10: Qualitative results from KNN, NILUT [ 6 ] , Fixed and U-Net [ 10 ] strategies trained on “NYC” dataset subset. KNN follows yellow-blue shift regions of GT while other solutions struggle.
Reference-based color grading aims to reproduce the tonal mood and lighting of a reference while preserving color harmony and scene structure. Existing photorealistic and filter-based methods often produce unstable tone mappings -- over-shifting or inconsistently retaining colors -- leading to unnatural results. We propose CanonCGT, a two-stage framework built on a canonical pivot -- a style-neutral intermediate representation for stable color mapping. The first stage canonicalizes the input by removing intrinsic tonal bias, and the second color-grades it to match the reference style. A dual-phase training scheme, DP-CGT, combines supervised preset learning with self-supervised refinement on unpaired photographs. CanonCGT delivers photorealistic and tonally consistent results across diverse datasets, surpassing state-of-the-art methods in stability and visual fidelity. Our codes are available at https://github.com/Jinwon-Ko/CanonCGT
Jinwon Ko, Keunsoo Ko, Chang-Su Kim
Korea University · The Catholic University of Korea
Personalized image enhancement should reflect individual aesthetic taste, yet learning such preferences commonly depends on private photos and ratings that are unsuitable for centralized collection. The task must infer preference from sparse, heterogeneous feedback and translate it into natural-looking color transformations on resource-constrained user devices. We introduce FedPAIE, a federated personalized aesthetic image enhancement framework for user-adaptive color grading without centralizing raw photos or ratings. FedPAIE trains a lightweight dual-cue aesthetic scorer, calibrates it into a personalized scorer on a small local support set, and freezes it to guide regularized adaptation of a lightweight CLUT enhancer from unpaired local photographs. Fidelity constraints and an excess-gap penalty regularize scorer-guided adaptation to limit proxy-score over-optimization while preserving content and natural appearance. Training remains lightweight throughout the pipeline: scorer learning updates at most 0.787M parameters, enhancer adaptation updates 0.265M, and inference retains only a 0.293M-parameter personalized enhancer. Experiments on MIT-Adobe FiveK and Flickr-AES demonstrate effective open-world personalization and a favorable balance between user preference and image fidelity. FedPAIE thus connects decentralized preference learning with efficient personalized image transformation without requiring paired user retouches.
Photorealistic style transfer aims to match the color and tone of an input image to that of a style target while preserving the content and details of the original scene. Although existing large image models can facilitate these kinds of appearance edits, their high computational demands, potential for hallucinations, and limited user control make them unsuitable for high-resolution, real-time workflows. We introduce Hist2Style, a bilateral-grid formulation for fast, edge-aware stylization that preserves visual fidelity by constraining operations to locally affine transforms in bilateral space. Our model distills a large image editing model into a lightweight network by training on a large supervised corpus generated with language and vision-language models, targeting spatially varying color edits. The network conditions on a histogram-based embedding of the style target to provide an interpretable interface for adjusting the output style by modifying the target color distribution. Overall, Hist2Style maintains content structure by construction, avoids hallucinations, and supports real-time, high-resolution photorealistic stylization with interactive user-controllable color and tone adjustments.
Dekel Galor, Adam Pikielny, Zhoutong Zhang +4
Adobe Nextcam · University of California, Berkeley