Seiichi Uchida

Prakash Chandra Chhipa, Richa Upadhyay, Gustav Grund Pihlgren et al.

This work presents a novel self-supervised pre-training method to learn efficient representations without labels on histopathology medical images utilizing magnification factors. Other state-of-theart works mainly focus on fully supervised learning approaches that rely heavily on human annotations. However, the scarcity of labeled and unlabeled data is a long-standing challenge in histopathology. Currently, representation learning without labels remains unexplored for the histopathology domain. The proposed method, Magnification Prior Contrastive Similarity (MPCS), enables self-supervised learning of representations without labels on small-scale breast cancer dataset BreakHis by exploiting magnification factor, inductive transfer, and reducing human prior. The proposed method matches fully supervised learning state-of-the-art performance in malignancy classification when only 20% of labels are used in fine-tuning and outperform previous works in fully supervised learning settings. It formulates a hypothesis and provides empirical evidence to support that reducing human-prior leads to efficient representation learning in self-supervision. The implementation of this work is available online on GitHub - https://github.com/prakashchhipa/Magnification-Prior-Self-Supervised-Method

Guangtao Lyu, Kun Liu, Anna Zhu et al.

The scene text removal (STR) task aims to remove text regions and recover the background smoothly in images for private information protection. Most existing STR methods adopt encoder-decoder-based CNNs, with direct copies of the features in the skip connections. However, the encoded features contain both text texture and structure information. The insufficient utilization of text features hampers the performance of background reconstruction in text removal regions. To tackle these problems, we propose a novel Feature Erasing and Transferring (FET) mechanism to reconfigure the encoded features for STR in this paper. In FET, a Feature Erasing Module (FEM) is designed to erase text features. An attention module is responsible for generating the feature similarity guidance. The Feature Transferring Module (FTM) is introduced to transfer the corresponding features in different layers based on the attention guidance. With this mechanism, a one-stage, end-to-end trainable network called FETNet is constructed for scene text removal. In addition, to facilitate research on both scene text removal and segmentation tasks, we introduce a novel dataset, Flickr-ST, with multi-category annotations. A sufficient number of experiments and ablation studies are conducted on the public datasets and Flickr-ST. Our proposed method achieves state-of-the-art performance using most metrics, with remarkably higher quality scene text removal results. The source code of our work is available at: \href{https://github.com/GuangtaoLyu/FETNet}{https://github.com/GuangtaoLyu/FETNet.

Prakash Chandra Chhipa, Richa Upadhyay, Rajkumar Saini et al.

This work presents a novel self-supervised representation learning method to learn efficient representations without labels on images from a 3DPM sensor (3-Dimensional Particle Measurement; estimates the particle size distribution of material) utilizing RGB images and depth maps of mining material on the conveyor belt. Human annotations for material categories on sensor-generated data are scarce and cost-intensive. Currently, representation learning without human annotations remains unexplored for mining materials and does not leverage on utilization of sensor-generated data. The proposed method, Depth Contrast, enables self-supervised learning of representations without labels on the 3DPM dataset by exploiting depth maps and inductive transfer. The proposed method outperforms material classification over ImageNet transfer learning performance in fully supervised learning settings and achieves an F1 score of 0.73. Further, The proposed method yields an F1 score of 0.65 with an 11% improvement over ImageNet transfer learning performance in a semi-supervised setting when only 20% of labels are used in fine-tuning. Finally, the Proposed method showcases improved performance generalization on linear evaluation. The implementation of proposed method is available on GitHub.

Shota Harada, Ryoma Bise, Kengo Araki et al.

Semi-supervised domain adaptation is a technique to build a classifier for a target domain by modifying a classifier in another (source) domain using many unlabeled samples and a small number of labeled samples from the target domain. In this paper, we develop a semi-supervised domain adaptation method, which has robustness to class-imbalanced situations, which are common in medical image classification tasks. For robustness, we propose a weakly-supervised clustering pipeline to obtain high-purity clusters and utilize the clusters in representation learning for domain adaptation. The proposed method showed state-of-the-art performance in the experiment using severely class-imbalanced pathological image patches.

Takahiro Shirakawa, Seiichi Uchida

Ambigrams are graphical letter designs that can be read not only from the original direction but also from a rotated direction (especially with 180 degrees). Designing ambigrams is difficult even for human experts because keeping their dual readability from both directions is often difficult. This paper proposes an ambigram generation model. As its generation module, we use a diffusion model, which has recently been used to generate high-quality photographic images. By specifying a pair of letter classes, such as 'A' and 'B', the proposed model generates various ambigram images which can be read as 'A' from the original direction and 'B' from a direction rotated 180 degrees. Quantitative and qualitative analyses of experimental results show that the proposed model can generate high-quality and diverse ambigrams. In addition, we define ambigramability, an objective measure of how easy it is to generate ambigrams for each letter pair. For example, the pair of 'A' and 'V' shows a high ambigramability (that is, it is easy to generate their ambigrams), and the pair of 'D' and 'K' shows a lower ambigramability. The ambigramability gives various hints of the ambigram generation not only for computers but also for human experts. The code can be found at (https://github.com/univ-esuty/ambifusion).

Shinnosuke Matsuo, Xiaomeng Wu, Gantugs Atarsaikhan et al.

Similarity measures for time series are important problems for time series classification. To handle the nonlinear time distortions, Dynamic Time Warping (DTW) has been widely used. However, DTW is not learnable and suffers from a trade-off between robustness against time distortion and discriminative power. In this paper, we propose a neural network model for task-adaptive time warping. Specifically, we use the attention model, called the bipartite attention model, to develop an explicit time warping mechanism with greater distortion invariance. Unlike other learnable models using DTW for warping, our model predicts all local correspondences between two time series and is trained based on metric learning, which enables it to learn the optimal data-dependent warping for the target task. We also propose to induce pre-training of our model by DTW to improve the discriminative power. Extensive experiments demonstrate the superior effectiveness of our model over DTW and its state-of-the-art performance in online signature verification.

Prakash Chandra Chhipa, Muskaan Chopra, Gopal Mengi et al.

This work investigates the unexplored usability of self-supervised representation learning in the direction of functional knowledge transfer. In this work, functional knowledge transfer is achieved by joint optimization of self-supervised learning pseudo task and supervised learning task, improving supervised learning task performance. Recent progress in self-supervised learning uses a large volume of data, which becomes a constraint for its applications on small-scale datasets. This work shares a simple yet effective joint training framework that reinforces human-supervised task learning by learning self-supervised representations just-in-time and vice versa. Experiments on three public datasets from different visual domains, Intel Image, CIFAR, and APTOS, reveal a consistent track of performance improvements on classification tasks during joint optimization. Qualitative analysis also supports the robustness of learnt representations. Source code and trained models are available on GitHub.

Masaya Ueda, Akisato Kimura, Seiichi Uchida

Different fonts have different impressions, such as elegant, scary, and cool. This paper tackles part-based shape-impression analysis based on the Transformer architecture, which is able to handle the correlation among local parts by its self-attention mechanism. This ability will reveal how combinations of local parts realize a specific impression of a font. The versatility of Transformer allows us to realize two very different approaches for the analysis, i.e., multi-label classification and translation. A quantitative evaluation shows that our Transformer-based approaches estimate the font impressions from a set of local parts more accurately than other approaches. A qualitative evaluation then indicates the important local parts for a specific impression.

Wataru Shimoda, Daichi Haraguchi, Seiichi Uchida et al.

In this work, we consider the typography generation task that aims at producing diverse typographic styling for the given graphic document. We formulate typography generation as a fine-grained attribute generation for multiple text elements and build an autoregressive model to generate diverse typography that matches the input design context. We further propose a simple yet effective sampling approach that respects the consistency and distinction principle of typography so that generated examples share consistent typographic styling across text elements. Our empirical study shows that our model successfully generates diverse typographic designs while preserving a consistent typographic structure.

Yusuke Nagata, Jinki Otao, Daichi Haraguchi et al.

We propose TrueType Transformer (T3), which can perform character and font style recognition in an outline format. The outline format, such as TrueType, represents each character as a sequence of control points of stroke contours and is frequently used in born-digital documents. T3 is organized by a deep neural network, so-called Transformer. Transformer is originally proposed for sequential data, such as text, and therefore appropriate for handling the outline data. In other words, T3 directly accepts the outline data without converting it into a bitmap image. Consequently, T3 realizes a resolution-independent classification. Moreover, since the locations of the control points represent the fine and local structures of the font style, T3 is suitable for font style classification, where such structures are very important. In this paper, we experimentally show the applicability of T3 in character and font style recognition tasks, while observing how the individual control points contribute to classification results.

Yusuke Nagata, Brian Kenji Iwana, Seiichi Uchida

In documents and graphics, contours are a popular format to describe specific shapes. For example, in the True Type Font (TTF) file format, contours describe vector outlines of typeface shapes. Each contour is often defined as a sequence of points. In this paper, we tackle the contour completion task. In this task, the input is a contour sequence with missing points, and the output is a generated completed contour. This task is more difficult than image completion because, for images, the missing pixels are indicated. Since there is no such indication in the contour completion task, we must solve the problem of missing part detection and completion simultaneously. We propose a Transformer-based method to solve this problem and show the results of the typeface contour completion.

Yuya Saito, Shinnosuke Matsuo, Seiichi Uchida et al.

This paper tackles the problem of the worst-class error rate, instead of the standard error rate averaged over all classes. For example, a three-class classification task with class-wise error rates of 10%, 10%, and 40% has a worst-class error rate of 40%, whereas the average is 20% under the class-balanced condition. The worst-class error is important in many applications. For example, in a medical image classification task, it would not be acceptable for the malignant tumor class to have a 40% error rate, while the benign and healthy classes have a 10% error rates. To avoid overfitting in worst-class error minimization using Deep Neural Networks (DNNs), we design a problem formulation for bounding the worst-class error instead of achieving zero worst-class error. Moreover, to correctly bound the worst-class error, we propose a boosting approach which ensembles DNNs. We give training and generalization worst-class-error bound. Experimental results show that the algorithm lowers worst-class test error rates while avoiding overfitting to the training set. This code is available at https://github.com/saito-yuya/Bounding-the-Worst-class-error-A-Boosting-Approach.

Seiya Matsuda, Akisato Kimura, Seiichi Uchida

Our goal is to generate fonts with specific impressions, by training a generative adversarial network with a font dataset with impression labels. The main difficulty is that font impression is ambiguous and the absence of an impression label does not always mean that the font does not have the impression. This paper proposes a font generation model that is robust against missing impression labels. The key ideas of the proposed method are (1)a co-occurrence-based missing label estimator and (2)an impression label space compressor. The first is to interpolate missing impression labels based on the co-occurrence of labels in the dataset and use them for training the model as completed label conditions. The second is an encoder-decoder module to compress the high-dimensional impression space into low-dimensional. We proved that the proposed model generates high-quality font images using multi-label data with missing labels through qualitative and quantitative evaluations.

Shumpei Takezaki, Kiyohito Tanaka, Seiichi Uchida et al.

Disease severity regression by a convolutional neural network (CNN) for medical images requires a sufficient number of image samples labeled with severity levels. Conditional generative adversarial network (cGAN)-based data augmentation (DA) is a possible solution, but it encounters two issues. The first issue is that existing cGANs cannot deal with real-valued severity levels as their conditions, and the second is that the severity of the generated images is not fully reliable. We propose continuous DA as a solution to the two issues. Our method uses continuous severity GAN to generate images at real-valued severity levels and dataset-disjoint multi-objective optimization to deal with the second issue. Our method was evaluated for estimating ulcerative colitis (UC) severity of endoscopic images and achieved higher classification performance than conventional DA methods.

Xiaotong Ji, Yuchen Zheng, Daiki Suehiro et al.

In this paper, we propose an optimal rejection method for rejecting ambiguous samples by a rejection function. This rejection function is trained together with a classification function under the framework of Learning-with-Rejection (LwR). The highlights of LwR are: (1) the rejection strategy is not heuristic but has a strong background from a machine learning theory, and (2) the rejection function can be trained on an arbitrary feature space which is different from the feature space for classification. The latter suggests we can choose a feature space that is more suitable for rejection. Although the past research on LwR focused only on its theoretical aspect, we propose to utilize LwR for practical pattern classification tasks. Moreover, we propose to use features from different CNN layers for classification and rejection. Our extensive experiments of notMNIST classification and character/non-character classification demonstrate that the proposed method achieves better performance than traditional rejection strategies.

Takeaki Kadota, Hideaki Hayashi, Ryoma Bise et al.

Automatic image-based disease severity estimation generally uses discrete (i.e., quantized) severity labels. Annotating discrete labels is often difficult due to the images with ambiguous severity. An easier alternative is to use relative annotation, which compares the severity level between image pairs. By using a learning-to-rank framework with relative annotation, we can train a neural network that estimates rank scores that are relative to severity levels. However, the relative annotation for all possible pairs is prohibitive, and therefore, appropriate sample pair selection is mandatory. This paper proposes a deep Bayesian active-learning-to-rank, which trains a Bayesian convolutional neural network while automatically selecting appropriate pairs for relative annotation. We confirmed the efficiency of the proposed method through experiments on endoscopic images of ulcerative colitis. In addition, we confirmed that our method is useful even with the severe class imbalance because of its ability to select samples from minor classes automatically.

Takahiro Shirakawa, Seiichi Uchida

Layout-aware text-to-image generation is a task to generate multi-object images that reflect layout conditions in addition to text conditions. The current layout-aware text-to-image diffusion models still have several issues, including mismatches between the text and layout conditions and quality degradation of generated images. This paper proposes a novel layout-aware text-to-image diffusion model called NoiseCollage to tackle these issues. During the denoising process, NoiseCollage independently estimates noises for individual objects and then crops and merges them into a single noise. This operation helps avoid condition mismatches; in other words, it can put the right objects in the right places. Qualitative and quantitative evaluations show that NoiseCollage outperforms several state-of-the-art models. These successful results indicate that the crop-and-merge operation of noises is a reasonable strategy to control image generation. We also show that NoiseCollage can be integrated with ControlNet to use edges, sketches, and pose skeletons as additional conditions. Experimental results show that this integration boosts the layout accuracy of ControlNet. The code is available at https://github.com/univ-esuty/noisecollage.

Takeaki Kadota, Hideaki Hayashi, Ryoma Bise et al.

Automatic image-based severity estimation is an important task in computer-aided diagnosis. Severity estimation by deep learning requires a large amount of training data to achieve a high performance. In general, severity estimation uses training data annotated with discrete (i.e., quantized) severity labels. Annotating discrete labels is often difficult in images with ambiguous severity, and the annotation cost is high. In contrast, relative annotation, in which the severity between a pair of images is compared, can avoid quantizing severity and thus makes it easier. We can estimate relative disease severity using a learning-to-rank framework with relative annotations, but relative annotation has the problem of the enormous number of pairs that can be annotated. Therefore, the selection of appropriate pairs is essential for relative annotation. In this paper, we propose a deep Bayesian active learning-to-rank that automatically selects appropriate pairs for relative annotation. Our method preferentially annotates unlabeled pairs with high learning efficiency from the model uncertainty of the samples. We prove the theoretical basis for adapting Bayesian neural networks to pairwise learning-to-rank and demonstrate the efficiency of our method through experiments on endoscopic images of ulcerative colitis on both private and public datasets. We also show that our method achieves a high performance under conditions of significant class imbalance because it automatically selects samples from the minority classes.

Naoya Yasukochi, Hideaki Hayashi, Daichi Haraguchi et al.

There are various font styles in the world. Different styles give different impressions and readability. This paper analyzes the relationship between font styles and contextual factors that might affect font style selection with large-scale datasets. For example, we will analyze the relationship between font style and its surrounding object (such as ``bus'') by using about 800,000 words in the Open Images dataset. We also use a book cover dataset to analyze the relationship between font styles with book genres. Moreover, the meaning of the word is assumed as another contextual factor. For these numeric analyses, we utilize our own font-style feature extraction model and word2vec. As a result of co-occurrence-based relationship analysis, we found several instances of specific font styles being used for specific contextual factors.

Phuong Ngoc Nguyen, Kaito Shiku, Ryoma Bise et al.

Active learning aims to reduce annotation cost by selectively querying informative samples for supervision under a limited labeling budget. In this work, we investigate how vision-language models (VLMs) can be leveraged to further reduce the reliance on costly human annotation within the active learning paradigm. To this end, we find that the reliability of VLMs varies significantly with label granularity in fine-grained recognition tasks: they perform poorly on fine-grained labels but can provide accurate coarse-grained labels. Leveraging this property, we propose an active learning framework that combines fine-grained human annotations with coarse-grained VLM-generated weak labels through instance-wise label assignment. We further model the systematic noise in VLM-generated labels using a small set of trusted full labels. Experiments on CUB200 and FGVC-Aircraft show that the proposed framework consistently outperforms existing active learning methods under the same annotation budget.

Shinnosuke Matsuo, Ryoma Bise, Seiichi Uchida et al.

This paper proposes a novel and efficient method for Learning from Label Proportions (LLP), whose goal is to train a classifier only by using the class label proportions of instance sets, called bags. We propose a novel LLP method based on an online pseudo-labeling method with regret minimization. As opposed to the previous LLP methods, the proposed method effectively works even if the bag sizes are large. We demonstrate the effectiveness of the proposed method using some benchmark datasets.

Shota Harada, Ryoma Bise, Kiyohito Tanaka et al.

Semi-supervised domain adaptation leverages a few labeled and many unlabeled target samples, making it promising for addressing domain shifts in medical image analysis. However, existing methods struggle with severity classification due to unclear class boundaries. Severity classification involves naturally ordered class labels, complicating adaptation. We propose a novel method that aligns source and target domains using rank scores learned via ranking with class order. Specifically, Cross-Domain Ranking ranks sample pairs across domains, while Continuous Distribution Alignment aligns rank score distributions. Experiments on ulcerative colitis and diabetic retinopathy classification validate the effectiveness of our approach, demonstrating successful alignment of class-specific rank score distributions.

Xiaotong Ji, Yan Zheng, Daiki Suehiro et al.

Signature verification, as a crucial practical documentation analysis task, has been continuously studied by researchers in machine learning and pattern recognition fields. In specific scenarios like confirming financial documents and legal instruments, ensuring the absolute reliability of signatures is of top priority. In this work, we proposed a new method to learn "top-rank pairs" for writer-independent offline signature verification tasks. By this scheme, it is possible to maximize the number of absolutely reliable signatures. More precisely, our method to learn top-rank pairs aims at pushing positive samples beyond negative samples, after pairing each of them with a genuine reference signature. In the experiment, BHSig-B and BHSig-H datasets are used for evaluation, on which the proposed model achieves overwhelming better pos@top (the ratio of absolute top positive samples to all of the positive samples) while showing encouraging performance on both Area Under the Curve (AUC) and accuracy.

Yuma Yamaoka, Seiichi Uchida, Shoji Toyota

A state-space model is a statistical framework for inferring latent states from observed time-series data. However, inference with nonlinear and high-dimensional state-space models remains challenging. To this end, an approach based on diffusion models-a powerful class of deep generative models-has been developed, known as Score-based Data Assimilation (SDA). However, SDA cannot be directly applied when the latent-state transition depends on unknown parameters that must be inferred jointly with the latent states. To overcome this limitation, we propose a framework that enables SDA to handle latent states with unknown parameters. A key feature of the proposed method is the incorporation of the self-organization technique, which has been used in classical state-space modeling for the joint estimation of latent states and parameters. By integrating this classical technique into modern SDA, our method enables joint inference of latent states and unknown parameters while maintaining the high training efficiency of SDA. The effectiveness of the proposed approach is validated through numerical experiments on dynamical systems arising in neuroscience and atmospheric science. In addition, its scalability is demonstrated using a high-dimensional Kolmogorov flow, with the data dimension on the order of several hundred thousand.

Shinnosuke Matsuo, Riku Togashi, Ryoma Bise et al.

Active learning (AL) is a label-efficient machine learning paradigm that focuses on selectively annotating high-value instances to maximize learning efficiency. Its effectiveness can be further enhanced by incorporating weak supervision, which uses rough yet cost-effective annotations instead of exact (i.e., full) but expensive annotations. We introduce a novel AL framework, Instance-wise Supervision-Level Optimization (ISO), which not only selects the instances to annotate but also determines their optimal annotation level within a fixed annotation budget. Its optimization criterion leverages the value-to-cost ratio (VCR) of each instance while ensuring diversity among the selected instances. In classification experiments, ISO consistently outperforms traditional AL methods and surpasses a state-of-the-art AL approach that combines full and weak supervision, achieving higher accuracy at a lower overall cost. This code is available at https://github.com/matsuo-shinnosuke/ISOAL.

Daichi Haraguchi, Seiichi Uchida

When we compare fonts, we often pay attention to styles of local parts, such as serifs and curvatures. This paper proposes an attention mechanism to find important local parts. The local parts with larger attention are then considered important. The proposed mechanism can be trained in a quasi-self-supervised manner that requires no manual annotation other than knowing that a set of character images is from the same font, such as Helvetica. After confirming that the trained attention mechanism can find style-relevant local parts, we utilize the resulting attention for local style-aware font generation. Specifically, we design a new reconstruction loss function to put more weight on the local parts with larger attention for generating character images with more accurate style realization. This loss function has the merit of applicability to various font generation models. Our experimental results show that the proposed loss function improves the quality of generated character images by several few-shot font generation models.

Chean Fei Shee, Seiichi Uchida

A multi-layer image is more valuable than a single-layer image from a graphic designer's perspective. However, most of the proposed image generation methods so far focus on single-layer images. In this paper, we propose MontageGAN, which is a Generative Adversarial Networks (GAN) framework for generating multi-layer images. Our method utilized a two-step approach consisting of local GANs and global GAN. Each local GAN learns to generate a specific image layer, and the global GAN learns the placement of each generated image layer. Through our experiments, we show the ability of our method to generate multi-layer images and estimate the placement of the generated image layers.

Shumpei Takezaki, Kiyohito Tanaka, Seiichi Uchida

Severity level estimation is a crucial task in medical image diagnosis. However, accurately assigning severity class labels to individual images is very costly and challenging. Consequently, the attached labels tend to be noisy. In this paper, we propose a new framework for training with ``ordinal'' noisy labels. Since severity levels have an ordinal relationship, we can leverage this to train a classifier while mitigating the negative effects of noisy labels. Our framework uses two techniques: clean sample selection and dual-network architecture. A technical highlight of our approach is the use of soft labels derived from noisy hard labels. By appropriately using the soft and hard labels in the two techniques, we achieve more accurate sample selection and robust network training. The proposed method outperforms various state-of-the-art methods in experiments using two endoscopic ulcerative colitis (UC) datasets and a retinal Diabetic Retinopathy (DR) dataset. Our codes are available at https://github.com/shumpei-takezaki/Self-Relaxed-Joint-Training.

Daichi Haraguchi, Wataru Shimoda, Kota Yamaguchi et al.

In this paper, we demonstrate a total disentanglement of font images. Total disentanglement is a neural network-based method for decomposing each font image nonlinearly and completely into its style and content (i.e., character class) features. It uses a simple but careful training procedure to extract the common style feature from all `A'-`Z' images in the same font and the common content feature from all `A' (or another class) images in different fonts. These disentangled features guarantee the reconstruction of the original font image. Various experiments have been conducted to understand the performance of total disentanglement. First, it is demonstrated that total disentanglement is achievable with very high accuracy; this is experimental proof of the long-standing open question, ``Does `A'-ness exist?'' Hofstadter (1985). Second, it is demonstrated that the disentangled features produced by total disentanglement apply to a variety of tasks, including font recognition, character recognition, and one-shot font image generation. Code is available here: https://github.com/uchidalab/total_disentanglement

Heon Song, Daiki Suehiro, Seiichi Uchida

For visual object tracking, it is difficult to realize an almighty online tracker due to the huge variations of target appearance depending on an image sequence. This paper proposes an online tracking method that adaptively aggregates arbitrary multiple online trackers. The performance of the proposed method is theoretically guaranteed to be comparable to that of the best tracker for any image sequence, although the best expert is unknown during tracking. The experimental study on the large variations of benchmark datasets and aggregated trackers demonstrates that the proposed method can achieve state-of-the-art performance. The code is available at https://github.com/songheony/AAA-journal.

Brian Kenji Iwana, Seiichi Uchida

Neural networks have become a powerful tool in pattern recognition and part of their success is due to generalization from using large datasets. However, unlike other domains, time series classification datasets are often small. In order to address this problem, we propose a novel time series data augmentation called guided warping. While many data augmentation methods are based on random transformations, guided warping exploits the element alignment properties of Dynamic Time Warping (DTW) and shapeDTW, a high-level DTW method based on shape descriptors, to deterministically warp sample patterns. In this way, the time series are mixed by warping the features of a sample pattern to match the time steps of a reference pattern. Furthermore, we introduce a discriminative teacher in order to serve as a directed reference for the guided warping. We evaluate the method on all 85 datasets in the 2015 UCR Time Series Archive with a deep convolutional neural network (CNN) and a recurrent neural network (RNN). The code with an easy to use implementation can be found at https://github.com/uchidalab/time_series_augmentation .

Tetta Kondo, Shumpei Takezaki, Daichi Haraguchi et al.

Fonts have huge variations in their styles and give readers different impressions. Therefore, generating new fonts is worthy of giving new impressions to readers. In this paper, we employ diffusion models to generate new font styles by interpolating a pair of reference fonts with different styles. More specifically, we propose three different interpolation approaches, image-blending, condition-blending, and noise-blending, with the diffusion models. We perform qualitative and quantitative experimental analyses to understand the style generation ability of the three approaches. According to experimental results, three proposed approaches can generate not only expected font styles but also somewhat serendipitous font styles. We also compare the approaches with a state-of-the-art style-conditional Latin-font generative network model to confirm the validity of using the diffusion models for the style interpolation task.

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 .

Shumpei Takezaki, Seiichi Uchida

Diffusion models have recently been used for medical image generation because of their high image quality. In this study, we focus on generating medical images with ordinal classes, which have ordinal relationships, such as severity levels. We propose an Ordinal Diffusion Model (ODM) that controls the ordinal relationships of the estimated noise images among the classes. Our model was evaluated experimentally by generating retinal and endoscopic images of multiple severity classes. ODM achieved higher performance than conventional generative models by generating realistic images, especially in high-severity classes with fewer training samples.

KhayTze Peong, Seiichi Uchida, Daichi Haraguchi

Recent diffusion-based generative models show promise in their ability to generate text images, but limitations in specifying the styles of the generated texts render them insufficient in the realm of typographic design. This paper proposes a typographic text generation system to add and modify text on typographic designs while specifying font styles, colors, and text effects. The proposed system is a novel combination of two off-the-shelf methods for diffusion models, ControlNet and Blended Latent Diffusion. The former functions to generate text images under the guidance of edge conditions specifying stroke contours. The latter blends latent noise in Latent Diffusion Models (LDM) to add typographic text naturally onto an existing background. We first show that given appropriate text edges, ControlNet can generate texts in specified fonts while incorporating effects described by prompts. We further introduce text edge manipulation as an intuitive and customizable way to produce texts with complex effects such as ``shadows'' and ``reflections''. Finally, with the proposed system, we successfully add and modify texts on a predefined background while preserving its overall coherence.

Shinnosuke Matsuo, Daiki Suehiro, Seiichi Uchida et al.

In this paper, we address the segmentation of tumor subtypes in whole slide images (WSI) by utilizing incomplete label proportions. Specifically, we utilize `partial' label proportions, which give the proportions among tumor subtypes but do not give the proportion between tumor and non-tumor. Partial label proportions are recorded as the standard diagnostic information by pathologists, and we, therefore, want to use them for realizing the segmentation model that can classify each WSI patch into one of the tumor subtypes or non-tumor. We call this problem ``learning from partial label proportions (LPLP)'' and formulate the problem as a weakly supervised learning problem. Then, we propose an efficient algorithm for this challenging problem by decomposing it into two weakly supervised learning subproblems: multiple instance learning (MIL) and learning from label proportions (LLP). These subproblems are optimized efficiently in the end-to-end manner. The effectiveness of our algorithm is demonstrated through experiments conducted on two WSI datasets.

Yuta Matsuzaki, Seiichi Uchida, Shumpei Takezaki

Diffusion models trained on noisy datasets often reproduce high-frequency training artifacts, significantly degrading generation quality. To address this, we propose SCoRe (Spectral Cutoff Regeneration), a training-free, generation-time spectral regeneration method for clean image generation from diffusion models trained on noisy images. Leveraging the spectral bias of diffusion models, which infer high-frequency details from low-frequency cues, SCoRe suppresses corrupted high-frequency components of a generated image via a frequency cutoff and regenerates them via SDEdit. Crucially, we derive a theoretical mapping between the cutoff frequency and the SDEdit initialization timestep based on Radially Averaged Power Spectral Density (RAPSD), which prevents excessive noise injection during regeneration. Experiments on synthetic (CIFAR-10) and real-world (SIDD) noisy datasets demonstrate that SCoRe substantially outperforms post-processing and noise-robust baselines, restoring samples closer to clean image distributions without any retraining or fine-tuning.

Daichi Haraguchi, Brian Kenji Iwana, Seiichi Uchida

This study analyzes the relationship between non-verbal information (e.g., genres) and text design (e.g., font style, character color, etc.) through the classification of book genres using text design on book covers. Text images have both semantic information about the word itself and other information (non-semantic information or visual design), such as font style, character color, etc. When we read a word printed on some materials, we receive impressions or other information from both the word itself and the visual design. Basically, we can understand verbal information only from semantic information, i.e., the words themselves; however, we can consider that text design is helpful for understanding other additional information (i.e., non-verbal information), such as impressions, genre, etc. To investigate the effect of text design, we analyze text design using words printed on book covers and their genres in two scenarios. First, we attempted to understand the importance of visual design for determining the genre (i.e., non-verbal information) of books by analyzing the differences in the relationship between semantic information/visual design and genres. In the experiment, we found that semantic information is sufficient to determine the genre; however, text design is helpful in adding more discriminative features for book genres. Second, we investigated the effect of each text design on book genres. As a result, we found that each text design characterizes some book genres. For example, font style is useful to add more discriminative features for genres of ``Mystery, Thriller \& Suspense'' and ``Christian books \& Bibles.''

Yugo Kubota, Daichi Haraguchi, Seiichi Uchida

Fonts convey different impressions to readers. These impressions often come from the font shapes. However, the correlation between fonts and their impression is weak and unstable because impressions are subjective. To capture such weak and unstable cross-modal correlation between font shapes and their impressions, we propose Impression-CLIP, which is a novel machine-learning model based on CLIP (Contrastive Language-Image Pre-training). By using the CLIP-based model, font image features and their impression features are pulled closer, and font image features and unrelated impression features are pushed apart. This procedure realizes co-embedding between font image and their impressions. In our experiment, we perform cross-modal retrieval between fonts and impressions through co-embedding. The results indicate that Impression-CLIP achieves better retrieval accuracy than the state-of-the-art method. Additionally, our model shows the robustness to noise and missing tags.

Yugo Kubota, Kaito Shiku, Seiichi Uchida

Font shapes can evoke a wide range of impressions, but the correspondence between fonts and impression descriptions is not one-to-one: some impressions are broadly compatible with diverse styles, whereas others strongly constrain the set of plausible fonts. We refer to this graded constraint strength as style specificity. In this paper, we propose a hyperbolic co-embedding framework that models font--impression correspondence through entailment rather than simple paired alignment. Font images and impression descriptions, represented as single tags or tag sets, are embedded in a shared hyperbolic space with two complementary entailment constraints: impression-to-font entailment and low-to-high style-specificity entailment among impressions. This formulation induces a radial structure in which low style-specificity impressions lie near the origin and high style-specificity impressions lie farther away, yielding an interpretable geometric measure of how strongly an impression constrains font style. Experiments on the MyFonts dataset demonstrate improved bidirectional retrieval over strong one-to-one baselines. In addition, traversal and tag-level analyses show that the learned space captures a coherent progression from ambiguous to more style-specific impressions and provides a meaningful, data-driven quantification of style specificity.

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.

Masaki Akiba, Shumpei Takezaki, Daichi Haraguchi et al.

This paper proposes a novel model of few-part-shot font generation, which designs an entire font based on a set of partial design elements, i.e., partial shapes. Unlike conventional few-shot font generation, which requires entire character shapes for a couple of character classes, our approach only needs partial shapes as input. The proposed model not only improves the efficiency of font creation but also provides insights into how partial design details influence the entire structure of the individual characters.

Yugo Kubota, Seiichi Uchida

Different font styles (i.e., font shapes) convey distinct impressions, indicating a close relationship between font shapes and word tags describing those impressions. This paper proposes a novel embedding method for impression tags that leverages these shape-impression relationships. For instance, our method assigns similar vectors to impression tags that frequently co-occur in order to represent impressions of fonts, whereas standard word embedding methods (e.g., BERT and CLIP) yield very different vectors. This property is particularly useful for impression-based font generation and font retrieval. Technically, we construct a graph whose nodes represent impression tags and whose edges encode co-occurrence relationships. Then, we apply spectral embedding to obtain the impression vectors for each tag. We compare our method with BERT and CLIP in qualitative and quantitative evaluations, demonstrating that our approach performs better in impression-guided font generation.

Xiaotong Ji, Ryoma Bise, Seiichi Uchida

In medical image processing, accurate diagnosis is of paramount importance. Leveraging machine learning techniques, particularly top-rank learning, shows significant promise by focusing on the most crucial instances. However, challenges arise from noisy labels and class-ambiguous instances, which can severely hinder the top-rank objective, as they may be erroneously placed among the top-ranked instances. To address these, we propose a novel approach that enhances toprank learning by integrating a rejection module. Cooptimized with the top-rank loss, this module identifies and mitigates the impact of outliers that hinder training effectiveness. The rejection module functions as an additional branch, assessing instances based on a rejection function that measures their deviation from the norm. Through experimental validation on a medical dataset, our methodology demonstrates its efficacy in detecting and mitigating outliers, improving the reliability and accuracy of medical image diagnoses.

Ryo Ishiyama, Shinnosuke Matsuo, Seiichi Uchida

Multi-stroke characters in scripts such as Chinese and Japanese can be highly complex, posing significant challenges for both native speakers and, especially, non-native learners. If these characters can be simplified without degrading their legibility, it could reduce learning barriers for non-native speakers, facilitate simpler and legible font designs, and contribute to efficient character-based communication systems. In this paper, we propose a framework to systematically simplify multi-stroke characters by selectively removing strokes while preserving their overall legibility. More specifically, we use a highly accurate character recognition model to assess legibility and remove those strokes that minimally impact it. Experimental results on 1,256 character classes with 5, 10, 15, and 20 strokes reveal several key findings, including the observation that even after removing multiple strokes, many characters remain distinguishable. These findings suggest the potential for more formalized simplification strategies.

Takumi Yoshimatsu, Shumpei Takezaki, Seiichi Uchida

Scene text removal (STR) aims to erase textual elements from images. It was originally intended for removing privacy-sensitiveor undesired texts from natural scene images, but is now also appliedto typographic images. STR typically detects text regions and theninpaints them. Although STR has advanced through neural networksand synthetic data, misuse risks have increased. This paper investi-gates Inverse STR (ISTR), which analyzes STR-processed images andfocuses on binary classification (detecting whether an image has un-dergone STR) and localizing removed text regions. We demonstrate inexperiments that these tasks are achievable with high accuracies, en-abling detection of potential misuse and improving STR. We also at-tempt to recover the removed text content by training a text recognizerto understand its difficulty.

Ryo Ishiyama, Takahiro Shirakawa, Seiichi Uchida et al.

We propose Test-Time Augmentation (TTA) as an effective technique for addressing combinatorial optimization problems, including the Traveling Salesperson Problem. In general, deep learning models possessing the property of invariance, where the output is uniquely determined regardless of the node indices, have been proposed to learn graph structures efficiently. In contrast, we interpret the permutation of node indices, which exchanges the elements of the distance matrix, as a TTA scheme. The results demonstrate that our method is capable of obtaining shorter solutions than the latest models. Furthermore, we show that the probability of finding a solution closer to an exact solution increases depending on the augmentation size.

Masahito Toba, Seiichi Uchida, Hideaki Hayashi

In pseudo-labeling (PL), which is a type of semi-supervised learning, pseudo-labels are assigned based on the confidence scores provided by the classifier; therefore, accurate confidence is important for successful PL. In this study, we propose a PL algorithm based on an energy-based model (EBM), which is referred to as the energy-based PL (EBPL). In EBPL, a neural network-based classifier and an EBM are jointly trained by sharing their feature extraction parts. This approach enables the model to learn both the class decision boundary and input data distribution, enhancing confidence calibration during network training. The experimental results demonstrate that EBPL outperforms the existing PL method in semi-supervised image classification tasks, with superior confidence calibration error and recognition accuracy.

Sho Shimotsumagari, Shumpei Takezaki, Daichi Haraguchi et al.

The purpose of this paper is to enable the conversion between machine-printed character images (i.e., font images) and handwritten character images through machine learning. For this purpose, we propose a novel unpaired image-to-image domain conversion method, CycleDM, which incorporates the concept of CycleGAN into the diffusion model. Specifically, CycleDM has two internal conversion models that bridge the denoising processes of two image domains. These conversion models are efficiently trained without explicit correspondence between the domains. By applying machine-printed and handwritten character images to the two modalities, CycleDM realizes the conversion between them. Our experiments for evaluating the converted images quantitatively and qualitatively found that ours performs better than other comparable approaches.