Line-Art Colorization¶
Image & 2D Art
Open Source¶
-
ColorizeDiffusionXL Discovered by AI
2026
An SDXL-based reference-based sketch colorization framework designed for high-resolution (1024px) outputs. It addresses color-geometry conflicts using a dual-branch architecture and Gram regularization loss to achieve high-fidelity character colorization and geometry disentanglement.
disentanglementhigh-resolutionsdxlsketch-colorization -
GapFill (No Pixel Left Behind: Filling Gaps in Anime Colorization) Discovered by AI
2026
A production-focused tool designed to automatically detect and fill tiny unpainted regions ("gaps") that frequently occur during the digital colorization of anime line art, significantly reducing manual cleanup time for artists.
animecolorizationgap-fillingproduction-tool -
InstanceAnimator: Multi-Instance Sketch Video Colorization Discovered by AI
2026
A Diffusion Transformer framework for multi-instance sketch video colorization that introduces Canvas Guidance and instance-level control to handle multi-character scenarios with high fidelity.
animationcolorizationdiffusion-transformermulti-instance -
OmniColor Discovered by AI
2026
A unified framework for multi-modal line art colorization that supports arbitrary combinations of control signals, including spatially-aligned conditions and semantic-reference conditions. It integrates these signals through specialized encoding strategies and an adaptive gating module to achieve precise, flexible, and temporally stable results.
control-signalsdiffusionmulti-modalstable-colorization -
TexSketch: Bringing Texture-Aware Colorization to Sketches Discovered by AI
2026
A controllable procedural framework for generating colored-sketch datasets with programmable artistic styles, enabling texture-aware sketch colorization.
colorizationproceduralsketchtexture-aware -
Uni-Animator: Towards Unified Visual Colorization Discovered by AI
2026
A unified Diffusion Transformer (DiT)-based framework for both image and video sketch colorization, eliminating the need for task-specific model adaptation.
animationcolorizationdiffusion-transformersketch -
AnimeColor: Reference-based Animation Colorization with Diffusion Transformers Discovered by AI
2025
A reference-based animation colorization framework leveraging Diffusion Transformers (DiT) to integrate sketch sequences into a video diffusion model for temporal consistency.
animationcolorizationdiffusion-transformerreference-color -
Cobra: Efficient Line Art COlorization with BRoAder References Discovered by AI
2025
An efficient and versatile reference-based line art colorization method that supports color hints and utilizes over 200 reference images while maintaining low latency and high consistency.
colorizationcomicdiffusionreference-color -
ColorizeDiffusion v2: Enhancing Reference-based Sketch Colorization Through Separating Utilities Discovered by AI
2025
An enhanced version of ColorizeDiffusion that addresses the training-inference distribution mismatch by separating reference representations, improving generalization to misaligned real-world sketches.
colorizationdiffusionreference-colorsketch -
DACoN: DINO for Anime Paint Bucket Colorization with Any Number of Reference Images Discovered by AI
2025
A framework that leverages vision foundation models (DINO) to capture part-level semantics in line drawings, enabling robust paint bucket colorization with any number of reference images.
colorizationdinoflat-colorpaint-bucket -
Follow-Your-Color: Multi-Instance Sketch Colorization Discovered by AI
2025
A diffusion-based framework for multi-instance sketch colorization that allows coloring individual objects and instances while maintaining color consistency across multiple characters.
colorizationdiffusionmulti-instancesketch -
GANime: Generating Anime and Manga Character Drawings from Sketches with Deep Learning Discovered by AI
2025
A deep learning framework that explores image-to-image translation models, specifically Conditional GANs (C-GANs), to automate the colorization of anime and manga character sketches. It generates high-resolution, fully colorized illustrations directly from raw line drawings.
animecolorizationgansketch -
LongAnimation: Long Animation Generation with Dynamic Global-Local Memory Discovered by AI
2025
A video generation framework designed for long-term animation line art colorization. It introduces a dynamic global-local memory mechanism to maintain global color consistency and character identity across long sequences, overcoming the limitations of local frame-overlapping paradigms.
animationcolorizationdiffusionvideo-colorization -
MangaDiT: Reference-Guided Line Art Colorization with Hierarchical Attention in Diffusion Transformers Discovered by AI
2025
A reference-guided line art colorization framework using Diffusion Transformers (DiT) that discovers semantic correspondences implicitly through hierarchical internal attention mechanisms.
colorizationdiffusion-transformermangareference-color -
MangaNinja: Line Art Colorization with Precise Reference Following Discovered by AI
2025
A reference-guided line art colorization method based on diffusion models. It introduces a patch shuffling module to align references with target line art and a point-driven control scheme for fine-grained color matching.
colorizationdiffusionmangareference-color -
Region-Wise Correspondence Prediction between Manga Line Art Images Discovered by AI
2025
A method for predicting region-wise correspondences between raw manga line art images, facilitating downstream tasks like line art colorization and in-between frame generation.
colorizationcorrespondencelineartmanga -
AniDoc: Animation Creation Made Easier Discovered by AI
2024
A video line art colorization tool based on video diffusion models that automatically converts sketch sequences into colored animations following a character design reference.
animationcolorizationreference-colorvideo-diffusion -
ColorizeDiffusion: Adjustable Sketch Colorization with Reference Image and Text Discovered by AI
2024
A Stable Diffusion-based sketch colorization framework that supports both reference images and text prompts, resolving conflicts between different conditioning modalities.
colorizationdiffusionreference-colorsketch -
Control Color Discovered by AI
2024
A multi-modal interactive image colorization method that leverages pre-trained Stable Diffusion models. It offers highly controllable colorization capabilities for line art and grayscale images using stroke, exemplar, and text prompt guidance.
controlnetinteractivemulti-modalstable-diffusion -
Learning Inclusion Matching for Animation Paint Bucket Colorization Discovered by AI
2024
A learning-based inclusion matching pipeline for animation paint bucket colorization that directs the network to comprehend inclusion relationships between line-enclosed segments.
animationcolorizationflat-colorpaint-bucket -
LineArt Discovered by AI
2024
A knowledge-guided, training-free appearance transfer framework for design drawings and line art. It leverages pre-trained ControlNet and IP-Adapter models to transfer complex textures and colors from reference images onto detailed line drawings while preserving structural accuracy.
appearance-transfercontrolnetip-adaptertraining-free -
LVCD Discovered by AI
2024
A video diffusion framework designed for reference-based line art video colorization. It utilizes a sketch-guided ControlNet, reference attention, and sequential sampling with an overlapped blending module to produce temporally consistent, high-quality colorized animation videos from line art drawings.
animationdiffusionreference-guidedvideo-colorization -
Paint Bucket Colorization Using Anime Character Color Design Sheets Discovered by AI
2024
A framework for keyframe and consecutive frame paint bucket colorization guided by RGB values from anime character color design sheets.
animationcolorizationflat-colorpaint-bucket -
Diffusart: Enhancing Line Art Colorization with Conditional Diffusion Models Discovered by AI
2023
A conditional diffusion-based framework for user-guided line art colorization that utilizes sparse color scribbles as hints to generate high-quality colored illustrations.
colorizationdiffusionscribblesketch -
FlatGAN: A Holistic Approach for Robust Flat-Coloring in High-Definition with Understanding Line Discontinuity Discovered by AI
2023
A GAN-based framework designed for robust flat-coloring of high-definition line art, capable of handling line discontinuities and segmenting regions for uniform coloring.
colorizationflat-colorgansketch -
PaintsTorch2: Self-Supervised User-Guided Anime Lineart Colorization with Style Control Discovered by AI
2021
A self-supervised, user-guided anime line art colorization system that allows interactive coloring using color hints and style control.
animecolorizationinteractivesketch -
Reference-Based Sketch Image Colorization using Augmented-Self Reference and Dense Semantic Correspondence Discovered by AI
2020
A reference-based sketch colorization method that uses an augmented-self reference training scheme and a Spatially Corresponding Feature Transfer (SCFT) module to establish dense semantic correspondence between the sketch and reference image.
colorizationreference-basedscftsketch -
Style2Paints Discovered by AI
2018
An AI-driven line art colorization tool that splits the colorization process into color planning (drafting) and rendering stages, allowing users to colorize line art with high-quality results and layered outputs.
colorizationflat-colorinteractive -
PaintsChainer Discovered by AI
2017
A classic, highly influential open-source line drawing colorization tool developed by Preferred Networks. It utilizes deep learning to automatically colorize sketches with optional user-provided color hints, laying the foundation for modern interactive AI colorization workflows.
chainercolorizationinteractivesketch -
Style2Paints Discovered by AI
2017
A pioneering AI-driven anime sketch colorization tool that allows users to colorize line art using style references and color hints, widely recognized for its high-quality and natural shading results.
animecolorizationgansketch
Proprietary / Industry Tools¶
-
Clip Studio Paint AI Colorize Discovered by AI
2018
Clip Studio Paint features an integrated AI-powered automatic colorization tool that allows digital artists to colorize line art using server-side deep learning. It supports automatic coloring as well as hint-based colorization where users can paint color splotches on a separate layer to guide the AI.
colorizationlineartproprietarysoftware -
Petalica Paint Discovered by AI
2017
A widely used AI-powered automatic line art colorization service developed by Preferred Networks and operated by pixiv, allowing artists to colorize sketches automatically or guide the process using color hints.
colorizationflat-colorinteractivesketch
