Kibeom Hong

Kibeom Hong, Seogkyu Jeon, Junsoo Lee et al.

To deliver the artistic expression of the target style, recent studies exploit the attention mechanism owing to its ability to map the local patches of the style image to the corresponding patches of the content image. However, because of the low semantic correspondence between arbitrary content and artworks, the attention module repeatedly abuses specific local patches from the style image, resulting in disharmonious and evident repetitive artifacts. To overcome this limitation and accomplish impeccable artistic style transfer, we focus on enhancing the attention mechanism and capturing the rhythm of patterns that organize the style. In this paper, we introduce a novel metric, namely pattern repeatability, that quantifies the repetition of patterns in the style image. Based on the pattern repeatability, we propose Aesthetic Pattern-Aware style transfer Networks (AesPA-Net) that discover the sweet spot of local and global style expressions. In addition, we propose a novel self-supervisory task to encourage the attention mechanism to learn precise and meaningful semantic correspondence. Lastly, we introduce the patch-wise style loss to transfer the elaborate rhythm of local patterns. Through qualitative and quantitative evaluations, we verify the reliability of the proposed pattern repeatability that aligns with human perception, and demonstrate the superiority of the proposed framework.

Seogkyu Jeon, Bei Liu, Pilhyeon Lee et al.

Training deep generative models usually requires a large amount of data. To alleviate the data collection cost, the task of zero-shot GAN adaptation aims to reuse well-trained generators to synthesize images of an unseen target domain without any further training samples. Due to the data absence, the textual description of the target domain and the vision-language models, e.g., CLIP, are utilized to effectively guide the generator. However, with only a single representative text feature instead of real images, the synthesized images gradually lose diversity as the model is optimized, which is also known as mode collapse. To tackle the problem, we propose a novel method to find semantic variations of the target text in the CLIP space. Specifically, we explore diverse semantic variations based on the informative text feature of the target domain while regularizing the uncontrolled deviation of the semantic information. With the obtained variations, we design a novel directional moment loss that matches the first and second moments of image and text direction distributions. Moreover, we introduce elastic weight consolidation and a relation consistency loss to effectively preserve valuable content information from the source domain, e.g., appearances. Through extensive experiments, we demonstrate the efficacy of the proposed methods in ensuring sample diversity in various scenarios of zero-shot GAN adaptation. We also conduct ablation studies to validate the effect of each proposed component. Notably, our model achieves a new state-of-the-art on zero-shot GAN adaptation in terms of both diversity and quality.

Mirae Do, Seogkyu Jeon, Pilhyeon Lee et al.

Domain adaptation for object detection (DAOD) has recently drawn much attention owing to its capability of detecting target objects without any annotations. To tackle the problem, previous works focus on aligning features extracted from partial levels (e.g., image-level, instance-level, RPN-level) in a two-stage detector via adversarial training. However, individual levels in the object detection pipeline are closely related to each other and this inter-level relation is unconsidered yet. To this end, we introduce a novel framework for DAOD with three proposed components: Multi-scale-aware Uncertainty Attention (MUA), Transferable Region Proposal Network (TRPN), and Dynamic Instance Sampling (DIS). With these modules, we seek to reduce the negative transfer effect during training while maximizing transferability as well as discriminability in both domains. Finally, our framework implicitly learns domain invariant regions for object detection via exploiting the transferable information and enhances the complementarity between different detection levels by collaboratively utilizing their domain information. Through ablation studies and experiments, we show that the proposed modules contribute to the performance improvement in a synergic way, demonstrating the effectiveness of our method. Moreover, our model achieves a new state-of-the-art performance on various benchmarks.

Namhyuk Ahn, Junsoo Lee, Chunggi Lee et al.

Recent progresses in large-scale text-to-image models have yielded remarkable accomplishments, finding various applications in art domain. However, expressing unique characteristics of an artwork (e.g. brushwork, colortone, or composition) with text prompts alone may encounter limitations due to the inherent constraints of verbal description. To this end, we introduce DreamStyler, a novel framework designed for artistic image synthesis, proficient in both text-to-image synthesis and style transfer. DreamStyler optimizes a multi-stage textual embedding with a context-aware text prompt, resulting in prominent image quality. In addition, with content and style guidance, DreamStyler exhibits flexibility to accommodate a range of style references. Experimental results demonstrate its superior performance across multiple scenarios, suggesting its promising potential in artistic product creation.

Seogkyu Jeon, Kibeom Hong, Hyeran Byun

Recent domain generalized semantic segmentation (DGSS) studies have achieved notable improvements by distilling semantic knowledge from Vision-Language Models (VLMs). However, they overlook the semantic misalignment between visual and textual contexts, which arises due to the rigidity of a fixed context prompt learned on a single source domain. To this end, we present a novel domain generalization framework for semantic segmentation, namely Domain-aware Prompt-driven Masked Transformer (DPMFormer). Firstly, we introduce domain-aware prompt learning to facilitate semantic alignment between visual and textual cues. To capture various domain-specific properties with a single source dataset, we propose domain-aware contrastive learning along with the texture perturbation that diversifies the observable domains. Lastly, to establish a framework resilient against diverse environmental changes, we have proposed the domain-robust consistency learning which guides the model to minimize discrepancies of prediction from original and the augmented images. Through experiments and analyses, we demonstrate the superiority of the proposed framework, which establishes a new state-of-the-art on various DGSS benchmarks. The code is available at https://github.com/jone1222/DPMFormer.

Namhyuk Ahn, Patrick Kwon, Jihye Back et al.

Cartoonization is a task that renders natural photos into cartoon styles. Previous deep cartoonization methods only have focused on end-to-end translation, which may hinder editability. Instead, we propose a novel solution with editing features of texture and color based on the cartoon creation process. To do that, we design a model architecture to have separate decoders, texture and color, to decouple these attributes. In the texture decoder, we propose a texture controller, which enables a user to control stroke style and abstraction to generate diverse cartoon textures. We also introduce an HSV color augmentation to induce the networks to generate diverse and controllable color translation. To the best of our knowledge, our work is the first deep approach to control the cartoonization at inference while showing profound quality improvement over to baselines.

Jaehoon Ahn, Sumin Cho, Harim Jung et al.

We introduce ConTEXTure, a generative network designed to create a texture map/atlas for a given 3D mesh using images from multiple viewpoints. The process begins with generating a front-view image from a text prompt, such as 'Napoleon, front view', describing the 3D mesh. Additional images from different viewpoints are derived from this front-view image and camera poses relative to it. ConTEXTure builds upon the TEXTure network, which uses text prompts for six viewpoints (e.g., 'Napoleon, front view', 'Napoleon, left view', etc.). However, TEXTure often generates images for non-front viewpoints that do not accurately represent those viewpoints.To address this issue, we employ Zero123++, which generates multiple view-consistent images for the six specified viewpoints simultaneously, conditioned on the initial front-view image and the depth maps of the mesh for the six viewpoints. By utilizing these view-consistent images, ConTEXTure learns the texture atlas from all viewpoint images concurrently, unlike previous methods that do so sequentially. This approach ensures that the rendered images from various viewpoints, including back, side, bottom, and top, are free from viewpoint irregularities.

Sungnyun Kim, Junsoo Lee, Kibeom Hong et al.

In this study, we aim to enhance the capabilities of diffusion-based text-to-image (T2I) generation models by integrating diverse modalities beyond textual descriptions within a unified framework. To this end, we categorize widely used conditional inputs into three modality types: structure, layout, and attribute. We propose a multimodal T2I diffusion model, which is capable of processing all three modalities within a single architecture without modifying the parameters of the pre-trained diffusion model, as only a small subset of components is updated. Our approach sets new benchmarks in multimodal generation through extensive quantitative and qualitative comparisons with existing conditional generation methods. We demonstrate that DiffBlender effectively integrates multiple sources of information and supports diverse applications in detailed image synthesis. The code and demo are available at https://github.com/sungnyun/diffblender.

Kunhee Kim, NaHyeon Park, Kibeom Hong et al.

Textual Inversion (TI) is an efficient approach to text-to-image personalization but often fails on complex prompts. We trace these failures to embedding norm inflation: learned tokens drift to out-of-distribution magnitudes, degrading prompt conditioning in pre-norm Transformers. Empirically, we show semantics are primarily encoded by direction in CLIP token space, while inflated norms harm contextualization; theoretically, we analyze how large magnitudes attenuate positional information and hinder residual updates in pre-norm blocks. We propose Directional Textual Inversion (DTI), which fixes the embedding magnitude to an in-distribution scale and optimizes only direction on the unit hypersphere via Riemannian SGD. We cast direction learning as MAP with a von Mises-Fisher prior, yielding a constant-direction prior gradient that is simple and efficient to incorporate. Across personalization tasks, DTI improves text fidelity over TI and TI-variants while maintaining subject similarity. Crucially, DTI's hyperspherical parameterization enables smooth, semantically coherent interpolation between learned concepts (slerp), a capability that is absent in standard TI. Our findings suggest that direction-only optimization is a robust and scalable path for prompt-faithful personalization.

Jungwon Seo, Ferhat Ozgur Catak, Chunming Rong et al.

Federated Learning (FL) enables privacy-preserving multi-source information fusion (MSIF) but is challenged by client drift in highly heterogeneous data settings. Many existing drift-mitigation strategies rely on reference-based techniques--such as gradient adjustments or proximal loss--that use historical snapshots (e.g., past gradients or previous global models) as reference points. When only a subset of clients participates in each training round, these historical references may not accurately capture the overall data distribution, leading to unstable training. In contrast, our proposed Gradient Centralized Federated Learning (GC-Fed) employs a hyperplane as a historically independent reference point to guide local training and enhance inter-client alignment. GC-Fed comprises two complementary components: Local GC, which centralizes gradients during local training, and Global GC, which centralizes updates during server aggregation. In our hybrid design, Local GC is applied to feature-extraction layers to harmonize client contributions, while Global GC refines classifier layers to stabilize round-wise performance. Theoretical analysis and extensive experiments on benchmark FL tasks demonstrate that GC-Fed effectively mitigates client drift and achieves up to a 20% improvement in accuracy under heterogeneous and partial participation conditions.

Seogkyu Jeon, Kibeom Hong, Pilhyeon Lee et al.

Domain generalization aims to enhance the model robustness against domain shift without accessing the target domain. Since the available source domains for training are limited, recent approaches focus on generating samples of novel domains. Nevertheless, they either struggle with the optimization problem when synthesizing abundant domains or cause the distortion of class semantics. To these ends, we propose a novel domain generalization framework where feature statistics are utilized for stylizing original features to ones with novel domain properties. To preserve class information during stylization, we first decompose features into high and low frequency components. Afterward, we stylize the low frequency components with the novel domain styles sampled from the manipulated statistics, while preserving the shape cues in high frequency ones. As the final step, we re-merge both components to synthesize novel domain features. To enhance domain robustness, we utilize the stylized features to maintain the model consistency in terms of features as well as outputs. We achieve the feature consistency with the proposed domain-aware supervised contrastive loss, which ensures domain invariance while increasing class discriminability. Experimental results demonstrate the effectiveness of the proposed feature stylization and the domain-aware contrastive loss. Through quantitative comparisons, we verify the lead of our method upon existing state-of-the-art methods on two benchmarks, PACS and Office-Home.

Kibeom Hong, Seogkyu Jeon, Huan Yang et al.

Style transfer aims to reproduce content images with the styles from reference images. Existing universal style transfer methods successfully deliver arbitrary styles to original images either in an artistic or a photo-realistic way. However, the range of 'arbitrary style' defined by existing works is bounded in the particular domain due to their structural limitation. Specifically, the degrees of content preservation and stylization are established according to a predefined target domain. As a result, both photo-realistic and artistic models have difficulty in performing the desired style transfer for the other domain. To overcome this limitation, we propose a unified architecture, Domain-aware Style Transfer Networks (DSTN) that transfer not only the style but also the property of domain (i.e., domainness) from a given reference image. To this end, we design a novel domainness indicator that captures the domainness value from the texture and structural features of reference images. Moreover, we introduce a unified framework with domain-aware skip connection to adaptively transfer the stroke and palette to the input contents guided by the domainness indicator. Our extensive experiments validate that our model produces better qualitative results and outperforms previous methods in terms of proxy metrics on both artistic and photo-realistic stylizations.

Seogkyu Jeon, Pilhyeon Lee, Kibeom Hong et al.

Face aging is the task aiming to translate the faces in input images to designated ages. To simplify the problem, previous methods have limited themselves only able to produce discrete age groups, each of which consists of ten years. Consequently, the exact ages of the translated results are unknown and it is unable to obtain the faces of different ages within groups. To this end, we propose the continuous face aging generative adversarial networks (CFA-GAN). Specifically, to make the continuous aging feasible, we propose to decompose image features into two orthogonal features: the identity and the age basis features. Moreover, we introduce the novel loss function for identity preservation which maximizes the cosine similarity between the original and the generated identity basis features. With the qualitative and quantitative evaluations on MORPH, we demonstrate the realistic and continuous aging ability of our model, validating its superiority against existing models. To the best of our knowledge, this work is the first attempt to handle continuous target ages.

Kibeom Hong, Youngjung Uh, Hyeran Byun

Training GANs on videos is even more sophisticated than on images because videos have a distinguished dimension: time. While recent methods designed a dedicated architecture considering time, generated videos are still far from indistinguishable from real videos. In this paper, we introduce ArrowGAN framework, where the discriminators learns to classify arrow of time as an auxiliary task and the generators tries to synthesize forward-running videos. We argue that the auxiliary task should be carefully chosen regarding the target domain. In addition, we explore categorical ArrowGAN with recent techniques in conditional image generation upon ArrowGAN framework, achieving the state-of-the-art performance on categorical video generation. Our extensive experiments validate the effectiveness of arrow of time as a self-supervisory task, and demonstrate that all our components of categorical ArrowGAN lead to the improvement regarding video inception score and Frechet video distance on three datasets: Weizmann, UCFsports, and UCF-101.