Naoto Inoue

Kotaro Kikuchi, Naoto Inoue, Mayu Otani et al.

Color is a critical design factor for web pages, affecting important factors such as viewer emotions and the overall trust and satisfaction of a website. Effective coloring requires design knowledge and expertise, but if this process could be automated through data-driven modeling, efficient exploration and alternative workflows would be possible. However, this direction remains underexplored due to the lack of a formalization of the web page colorization problem, datasets, and evaluation protocols. In this work, we propose a new dataset consisting of e-commerce mobile web pages in a tractable format, which are created by simplifying the pages and extracting canonical color styles with a common web browser. The web page colorization problem is then formalized as a task of estimating plausible color styles for a given web page content with a given hierarchical structure of the elements. We present several Transformer-based methods that are adapted to this task by prepending structural message passing to capture hierarchical relationships between elements. Experimental results, including a quantitative evaluation designed for this task, demonstrate the advantages of our methods over statistical and image colorization methods. The code is available at https://github.com/CyberAgentAILab/webcolor.

Naoto Inoue, Kotaro Kikuchi, Edgar Simo-Serra et al.

Controllable layout generation aims at synthesizing plausible arrangement of element bounding boxes with optional constraints, such as type or position of a specific element. In this work, we try to solve a broad range of layout generation tasks in a single model that is based on discrete state-space diffusion models. Our model, named LayoutDM, naturally handles the structured layout data in the discrete representation and learns to progressively infer a noiseless layout from the initial input, where we model the layout corruption process by modality-wise discrete diffusion. For conditional generation, we propose to inject layout constraints in the form of masking or logit adjustment during inference. We show in the experiments that our LayoutDM successfully generates high-quality layouts and outperforms both task-specific and task-agnostic baselines on several layout tasks.

Naoto Inoue, Kotaro Kikuchi, Edgar Simo-Serra et al.

Creative workflows for generating graphical documents involve complex inter-related tasks, such as aligning elements, choosing appropriate fonts, or employing aesthetically harmonious colors. In this work, we attempt at building a holistic model that can jointly solve many different design tasks. Our model, which we denote by FlexDM, treats vector graphic documents as a set of multi-modal elements, and learns to predict masked fields such as element type, position, styling attributes, image, or text, using a unified architecture. Through the use of explicit multi-task learning and in-domain pre-training, our model can better capture the multi-modal relationships among the different document fields. Experimental results corroborate that our single FlexDM is able to successfully solve a multitude of different design tasks, while achieving performance that is competitive with task-specific and costly baselines.

Kotaro Kikuchi, Ukyo Honda, Naoto Inoue et al.

We introduce MarkupDM, a multimodal markup document model that represents graphic design as an interleaved multimodal document consisting of both markup language and images. Unlike existing holistic approaches that rely on an element-by-attribute grid representation, our representation accommodates variable-length elements, type-dependent attributes, and text content. Inspired by fill-in-the-middle training in code generation, we train the model to complete the missing part of a design document from its surrounding context, allowing it to treat various design tasks in a unified manner. Our model also supports image generation by predicting discrete image tokens through a specialized tokenizer with support for image transparency. We evaluate MarkupDM on three tasks, attribute value, image, and text completion, and demonstrate that it can produce plausible designs consistent with the given context. To further illustrate the flexibility of our approach, we evaluate our approach on a new instruction-guided design completion task where our instruction-tuned MarkupDM compares favorably to state-of-the-art image editing models, especially in textual completion. These findings suggest that multimodal language models with our document representation can serve as a versatile foundation for broad design automation.

Daichi Horita, Naoto Inoue, Kotaro Kikuchi et al.

Content-aware graphic layout generation aims to automatically arrange visual elements along with a given content, such as an e-commerce product image. In this paper, we argue that the current layout generation approaches suffer from the limited training data for the high-dimensional layout structure. We show that a simple retrieval augmentation can significantly improve the generation quality. Our model, which is named Retrieval-Augmented Layout Transformer (RALF), retrieves nearest neighbor layout examples based on an input image and feeds these results into an autoregressive generator. Our model can apply retrieval augmentation to various controllable generation tasks and yield high-quality layouts within a unified architecture. Our extensive experiments show that RALF successfully generates content-aware layouts in both constrained and unconstrained settings and significantly outperforms the baselines.

Mayu Otani, Naoto Inoue, Kotaro Kikuchi et al.

We introduce a layout similarity measure designed to evaluate the results of layout generation. While several similarity measures have been proposed in prior research, there has been a lack of comprehensive discussion about their behaviors. Our research uncovers that the majority of these measures are unable to handle various layout differences, primarily due to their dependencies on strict element matching, that is one-by-one matching of elements within the same category. To overcome this limitation, we propose a new similarity measure based on optimal transport, which facilitates a more flexible matching of elements. This approach allows us to quantify the similarity between any two layouts even those sharing no element categories, making our measure highly applicable to a wide range of layout generation tasks. For tasks such as unconditional layout generation, where FID is commonly used, we also extend our measure to deal with collection-level similarities between groups of layouts. The empirical result suggests that our collection-level measure offers more reliable comparisons than existing ones like FID and Max.IoU.

Shunki Uebayashi, Kento Masui, Kyohei Atarashi et al.

Multimodal Large Language Models (MLLMs) have recently emerged as general architectures capable of reasoning over diverse modalities. Benchmarks for MLLMs should measure their ability for cross-modal integration. However, current benchmarks are filled with shortcut questions, which can be solved using only a single modality, thereby yielding unreliable rankings. For example, in vision-language cases, we can find the correct answer without either the image or the text. These low-quality questions unnecessarily increase the size and computational requirements of benchmarks. We introduce a multi-modal and multidimensional item response theory framework (M3IRT) that extends classical IRT by decomposing both model ability and item difficulty into image-only, text-only, and cross-modal components. M3IRT estimates cross-modal ability of MLLMs and each question's cross-modal difficulty, enabling compact, high-quality subsets that better reflect multimodal reasoning. Across 24 VLMs on three benchmarks, M3IRT prioritizes genuinely cross-modal questions over shortcuts and preserves ranking fidelity even when 50% of items are artificially generated low-quality questions, thereby reducing evaluation cost while improving reliability. M3IRT thus offers a practical tool for assessing cross-modal reasoning and refining multimodal benchmarks.

Masayuki Kawarada, Kosuke Yamada, Antonio Tejero-de-Pablos et al.

Conditional image embeddings are feature representations that focus on specific aspects of an image indicated by a given textual condition (e.g., color, genre), which has been a challenging problem. Although recent vision foundation models, such as CLIP, offer rich representations of images, they are not designed to focus on a specified condition. In this paper, we propose DIOR, a method that leverages a large vision-language model (LVLM) to generate conditional image embeddings. DIOR is a training-free approach that prompts the LVLM to describe an image with a single word related to a given condition. The hidden state vector of the LVLM's last token is then extracted as the conditional image embedding. DIOR provides a versatile solution that can be applied to any image and condition without additional training or task-specific priors. Comprehensive experimental results on conditional image similarity tasks demonstrate that DIOR outperforms existing training-free baselines, including CLIP. Furthermore, DIOR achieves superior performance compared to methods that require additional training across multiple settings.

Naoto Inoue, Toshihiko Yamasaki

Shadow removal is an essential task in computer vision and computer graphics. Recent shadow removal approaches all train convolutional neural networks (CNN) on real paired shadow/shadow-free or shadow/shadow-free/mask image datasets. However, obtaining a large-scale, diverse, and accurate dataset has been a big challenge, and it limits the performance of the learned models on shadow images with unseen shapes/intensities. To overcome this challenge, we present SynShadow, a novel large-scale synthetic shadow/shadow-free/matte image triplets dataset and a pipeline to synthesize it. We extend a physically-grounded shadow illumination model and synthesize a shadow image given an arbitrary combination of a shadow-free image, a matte image, and shadow attenuation parameters. Owing to the diversity, quantity, and quality of SynShadow, we demonstrate that shadow removal models trained on SynShadow perform well in removing shadows with diverse shapes and intensities on some challenging benchmarks. Furthermore, we show that merely fine-tuning from a SynShadow-pre-trained model improves existing shadow detection and removal models. Codes are publicly available at https://github.com/naoto0804/SynShadow.

Ryosuke Furuta, Naoto Inoue, Toshihiko Yamasaki

This paper tackles a new problem setting: reinforcement learning with pixel-wise rewards (pixelRL) for image processing. After the introduction of the deep Q-network, deep RL has been achieving great success. However, the applications of deep reinforcement learning (RL) for image processing are still limited. Therefore, we extend deep RL to pixelRL for various image processing applications. In pixelRL, each pixel has an agent, and the agent changes the pixel value by taking an action. We also propose an effective learning method for pixelRL that significantly improves the performance by considering not only the future states of the own pixel but also those of the neighbor pixels. The proposed method can be applied to some image processing tasks that require pixel-wise manipulations, where deep RL has never been applied. Besides, it is possible to visualize what kind of operation is employed for each pixel at each iteration, which would help us understand why and how such an operation is chosen. We also believe that our technology can enhance the explainability and interpretability of the deep neural networks. In addition, because the operations executed at each pixels are visualized, we can change or modify the operations if necessary. We apply the proposed method to a variety of image processing tasks: image denoising, image restoration, local color enhancement, and saliency-driven image editing. Our experimental results demonstrate that the proposed method achieves comparable or better performance, compared with the state-of-the-art methods based on supervised learning. The source code is available on https://github.com/rfuruta/pixelRL.

Julian Jorge Andrade Guerreiro, Naoto Inoue, Kento Masui et al.

Finding a suitable layout represents a crucial task for diverse applications in graphic design. Motivated by simpler and smoother sampling trajectories, we explore the use of Flow Matching as an alternative to current diffusion-based layout generation models. Specifically, we propose LayoutFlow, an efficient flow-based model capable of generating high-quality layouts. Instead of progressively denoising the elements of a noisy layout, our method learns to gradually move, or flow, the elements of an initial sample until it reaches its final prediction. In addition, we employ a conditioning scheme that allows us to handle various generation tasks with varying degrees of conditioning with a single model. Empirically, LayoutFlow performs on par with state-of-the-art models while being significantly faster.

Daichi Haraguchi, Naoto Inoue, Wataru Shimoda et al.

Recent advancements in foundation models show promising capability in graphic design generation. Several studies have started employing Large Multimodal Models (LMMs) to evaluate graphic designs, assuming that LMMs can properly assess their quality, but it is unclear if the evaluation is reliable. One way to evaluate the quality of graphic design is to assess whether the design adheres to fundamental graphic design principles, which are the designer's common practice. In this paper, we compare the behavior of GPT-based evaluation and heuristic evaluation based on design principles using human annotations collected from 60 subjects. Our experiments reveal that, while GPTs cannot distinguish small details, they have a reasonably good correlation with human annotation and exhibit a similar tendency to heuristic metrics based on design principles, suggesting that they are indeed capable of assessing the quality of graphic design. Our dataset is available at https://cyberagentailab.github.io/Graphic-design-evaluation .

Tomoyuki Suzuki, Kang-Jun Liu, Naoto Inoue et al.

Designers craft and edit graphic designs in a layer representation, but layer-based editing becomes impossible once composited into a raster image. In this work, we propose LayerD, a method to decompose raster graphic designs into layers for re-editable creative workflow. LayerD addresses the decomposition task by iteratively extracting unoccluded foreground layers. We propose a simple yet effective refinement approach taking advantage of the assumption that layers often exhibit uniform appearance in graphic designs. As decomposition is ill-posed and the ground-truth layer structure may not be reliable, we develop a quality metric that addresses the difficulty. In experiments, we show that LayerD successfully achieves high-quality decomposition and outperforms baselines. We also demonstrate the use of LayerD with state-of-the-art image generators and layer-based editing.

Wataru Shimoda, Naoto Inoue, Daichi Haraguchi et al.

While recent text-to-image models can generate photorealistic images from text prompts that reflect detailed instructions, they still face significant challenges in accurately rendering words in the image. In this paper, we propose to retouch erroneous text renderings in the post-processing pipeline. Our approach, called Type-R, identifies typographical errors in the generated image, erases the erroneous text, regenerates text boxes for missing words, and finally corrects typos in the rendered words. Through extensive experiments, we show that Type-R, in combination with the latest text-to-image models such as Stable Diffusion or Flux, achieves the highest text rendering accuracy while maintaining image quality and also outperforms text-focused generation baselines in terms of balancing text accuracy and image quality.

Jun Muraoka, Daichi Haraguchi, Naoto Inoue et al.

In layout design for advertisements and web pages, balancing visual appeal and communication efficiency is crucial. This study examines automated text box placement in incomplete layouts, comparing a standard Transformer-based method, a small Vision and Language Model (Phi3.5-vision), a large pretrained VLM (Gemini), and an extended Transformer that processes multiple images. Evaluations on the Crello dataset show the standard Transformer-based models generally outperform VLM-based approaches, particularly when incorporating richer appearance information. However, all methods face challenges with very small text or densely populated layouts. These findings highlight the benefits of task-specific architectures and suggest avenues for further improvement in automated layout design.

Ding Xia, Naoto Inoue, Qianru Qiu et al.

Colors play a crucial role in the design of vector graphic documents by enhancing visual appeal, facilitating communication, improving usability, and ensuring accessibility. In this context, color recommendation involves suggesting appropriate colors to complete or refine a design when one or more colors are missing or require alteration. Traditional methods often struggled with these challenges due to the complex nature of color design and the limited data availability. In this study, we explored the use of pretrained Large Language Models (LLMs) and their commonsense reasoning capabilities for color recommendation, raising the question: Can pretrained LLMs serve as superior designers for color recommendation tasks? To investigate this, we developed a robust, rigorously validated pipeline, ColorGPT, that was built by systematically testing multiple color representations and applying effective prompt engineering techniques. Our approach primarily targeted color palette completion by recommending colors based on a set of given colors and accompanying context. Moreover, our method can be extended to full palette generation, producing an entire color palette corresponding to a provided textual description. Experimental results demonstrated that our LLM-based pipeline outperformed existing methods in terms of color suggestion accuracy and the distribution of colors in the color palette completion task. For the full palette generation task, our approach also yielded improvements in color diversity and similarity compared to current techniques.

Naoto Inoue, Kento Masui, Wataru Shimoda et al.

Automatic generation of graphic designs has recently received considerable attention. However, the state-of-the-art approaches are complex and rely on proprietary datasets, which creates reproducibility barriers. In this paper, we propose an open framework for automatic graphic design called OpenCOLE, where we build a modified version of the pioneering COLE and train our model exclusively on publicly available datasets. Based on GPT4V evaluations, our model shows promising performance comparable to the original COLE. We release the pipeline and training results to encourage open development.

Takehiko Ohkawa, Naoto Inoue, Hirokatsu Kataoka et al.

Unpaired image-to-image (I2I) translation has received considerable attention in pattern recognition and computer vision because of recent advancements in generative adversarial networks (GANs). However, due to the lack of explicit supervision, unpaired I2I models often fail to generate realistic images, especially in challenging datasets with different backgrounds and poses. Hence, stabilization is indispensable for GANs and applications of I2I translation. Herein, we propose Augmented Cyclic Consistency Regularization (ACCR), a novel regularization method for unpaired I2I translation. Our main idea is to enforce consistency regularization originating from semi-supervised learning on the discriminators leveraging real, fake, reconstructed, and augmented samples. We regularize the discriminators to output similar predictions when fed pairs of original and perturbed images. We qualitatively clarify why consistency regularization on fake and reconstructed samples works well. Quantitatively, our method outperforms the consistency regularized GAN (CR-GAN) in real-world translations and demonstrates efficacy against several data augmentation variants and cycle-consistent constraints.

Ryosuke Furuta, Naoto Inoue, Toshihiko Yamasaki

This paper tackles a new problem setting: reinforcement learning with pixel-wise rewards (pixelRL) for image processing. After the introduction of the deep Q-network, deep RL has been achieving great success. However, the applications of deep RL for image processing are still limited. Therefore, we extend deep RL to pixelRL for various image processing applications. In pixelRL, each pixel has an agent, and the agent changes the pixel value by taking an action. We also propose an effective learning method for pixelRL that significantly improves the performance by considering not only the future states of the own pixel but also those of the neighbor pixels. The proposed method can be applied to some image processing tasks that require pixel-wise manipulations, where deep RL has never been applied. We apply the proposed method to three image processing tasks: image denoising, image restoration, and local color enhancement. Our experimental results demonstrate that the proposed method achieves comparable or better performance, compared with the state-of-the-art methods based on supervised learning.

Naoto Inoue, Ryosuke Furuta, Toshihiko Yamasaki et al.

Can we detect common objects in a variety of image domains without instance-level annotations? In this paper, we present a framework for a novel task, cross-domain weakly supervised object detection, which addresses this question. For this paper, we have access to images with instance-level annotations in a source domain (e.g., natural image) and images with image-level annotations in a target domain (e.g., watercolor). In addition, the classes to be detected in the target domain are all or a subset of those in the source domain. Starting from a fully supervised object detector, which is pre-trained on the source domain, we propose a two-step progressive domain adaptation technique by fine-tuning the detector on two types of artificially and automatically generated samples. We test our methods on our newly collected datasets containing three image domains, and achieve an improvement of approximately 5 to 20 percentage points in terms of mean average precision (mAP) compared to the best-performing baselines.