Boeun Kim
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Boeun Kim has a total of 40 co-authors. Here are the top co-authors by number of shared publications:
- Hyung Jin Chang
University of Birmingham
5 shared publications Jungho Kim
Korea Electronics Technology Institute
4 shared publicationsHea In Jeong
Korea Electronics Technology Institute
3 shared publicationsHyung Jin Chang
University of Birmingham
3 shared publicationsSaim Shin
Korea Electronics Technology Institute
3 shared publicationsChung-Il Kim
Korea Electronics Technology Institute (KETI),Artificial Intelligence Research Center,South Korea
2 shared publications
Latest Publications
- 25 March 2026
- Published by Elsevier in Pattern Recognition
- Vol. 179
Bidirectional regression for monocular 6DoF head pose estimation and reference system alignment
- 25 March 2026
- Published by Elsevier in Pattern Recognition
- Vol. 179
Bidirectional regression for monocular 6DoF head pose estimation and reference system alignment
- Open Access
- 27 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 10516-10524
Roll Your Eyes: Gaze Redirection via Explicit 3D Eyeball Rotation
- 27 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 10516-10524
Open AccessRoll Your Eyes: Gaze Redirection via Explicit 3D Eyeball Rotation
- 10 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 22756-22765
PersonaBooth: Personalized Text-to-Motion Generation
This paper introduces Motion Personalization, a new task that generates personalized motions aligned with text descriptions using several basic motions containing Persona. To support this novel task, we introduce a new large-scale motion dataset called PerMo (PersonaMotion), which captures the unique personas of multiple actors. We also propose a multi-modal finetuning method of a pretrained motion diffusion model called PersonaBooth. PersonaBooth addresses two main challenges: i) A significant distribution gap between the persona-focused PerMo dataset and the pretraining datasets, which lack persona-specific data, and ii) the difficulty of capturing a consistent persona from the motions vary in content (action type). To tackle the dataset distribution gap, we introduce a persona token to accept new persona features and perform multi-modal adaptation for both text and visuals during finetuning. To capture a consistent persona, we incorporate a contrastive learning technique to enhance intra-cohesion among samples with the same persona. Furthermore, we introduce a context-aware fusion mechanism to maximize the integration of persona cues from multiple input motions. PersonaBooth outperforms state-of-the-art motion style transfer methods, establishing a new benchmark for motion personalization.
- 10 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 22756-22765
PersonaBooth: Personalized Text-to-Motion Generation
This paper introduces Motion Personalization, a new task that generates personalized motions aligned with text descriptions using several basic motions containing Persona. To support this novel task, we introduce a new large-scale motion dataset called PerMo (PersonaMotion), which captures the unique personas of multiple actors. We also propose a multi-modal finetuning method of a pretrained motion diffusion model called PersonaBooth. PersonaBooth addresses two main challenges: i) A significant distribution gap between the persona-focused PerMo dataset and the pretraining datasets, which lack persona-specific data, and ii) the difficulty of capturing a consistent persona from the motions vary in content (action type). To tackle the dataset distribution gap, we introduce a persona token to accept new persona features and perform multi-modal adaptation for both text and visuals during finetuning. To capture a consistent persona, we incorporate a contrastive learning technique to enhance intra-cohesion among samples with the same persona. Furthermore, we introduce a context-aware fusion mechanism to maximize the integration of persona cues from multiple input motions. PersonaBooth outperforms state-of-the-art motion style transfer methods, establishing a new benchmark for motion personalization.
- 10 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 23891-23900
3D Prior is All You Need: Cross-Task Few-shot 2D Gaze Estimation
3D and 2D gaze estimation share the fundamental objective of capturing eye movements but are traditionally treated as two distinct research domains. In this paper, we introduce a novel cross-task few-shot 2D gaze estimation approach, aiming to adapt a pre-trained 3D gaze estimation network for 2D gaze prediction on unseen devices using only a few training images. This task is highly challenging due to the domain gap between 3D and 2D gaze, unknown screen poses, and limited training data. To address these challenges, we propose a novel framework that bridges the gap between 3D and 2D gaze. Our framework contains a physics-based differentiable projection module with learnable parameters to model screen poses and project 3D gaze into 2D gaze. The framework is fully differentiable and can integrate into existing 3D gaze networks without modifying their original architecture. Additionally, we introduce a dynamic pseudo-labelling strategy for flipped images, which is particularly challenging for 2D labels due to unknown screen poses. To overcome this, we reverse the projection process by converting 2D labels to 3D space, where flipping is performed. Notably, this 3D space is not aligned with the camera coordinate system, so we learn a dynamic transformation matrix to compensate for this misalignment. We evaluate our method on MPIIGaze, EVE, and GazeCapture datasets, collected respectively on laptops, desktop computers, and mobile devices. The superior performance highlights the effectiveness of our approach, and demonstrates its strong potential for real-world applications.
- 10 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 23891-23900
3D Prior is All You Need: Cross-Task Few-shot 2D Gaze Estimation
3D and 2D gaze estimation share the fundamental objective of capturing eye movements but are traditionally treated as two distinct research domains. In this paper, we introduce a novel cross-task few-shot 2D gaze estimation approach, aiming to adapt a pre-trained 3D gaze estimation network for 2D gaze prediction on unseen devices using only a few training images. This task is highly challenging due to the domain gap between 3D and 2D gaze, unknown screen poses, and limited training data. To address these challenges, we propose a novel framework that bridges the gap between 3D and 2D gaze. Our framework contains a physics-based differentiable projection module with learnable parameters to model screen poses and project 3D gaze into 2D gaze. The framework is fully differentiable and can integrate into existing 3D gaze networks without modifying their original architecture. Additionally, we introduce a dynamic pseudo-labelling strategy for flipped images, which is particularly challenging for 2D labels due to unknown screen poses. To overcome this, we reverse the projection process by converting 2D labels to 3D space, where flipping is performed. Notably, this 3D space is not aligned with the camera coordinate system, so we learn a dynamic transformation matrix to compensate for this misalignment. We evaluate our method on MPIIGaze, EVE, and GazeCapture datasets, collected respectively on laptops, desktop computers, and mobile devices. The superior performance highlights the effectiveness of our approach, and demonstrates its strong potential for real-world applications.
- Open Access
- 10 November 2024
- Published by MDPI AG in Applied Sciences
- Vol. 14 (22) , 10322
Supervised Contrastive Learning for 3D Cross-Modal Retrieval
Interoperability between different virtual platforms requires the ability to search and transfer digital assets across platforms. Digital assets in virtual platforms are represented in different forms or modalities, such as images, meshes, and point clouds. The cross-modal retrieval of three-dimensional (3D) object representations is challenging due to data representation diversity, making common feature space discovery difficult. Recent studies have been focused on obtaining feature consistency within the same classes and modalities using cross-modal center loss. However, center features are sensitive to hyperparameter variations, making cross-modal center loss susceptible to performance degradation. This paper proposes a new 3D cross-modal retrieval method that uses cross-modal supervised contrastive learning (CSupCon) and the fixed projection head (FPH) strategy. Contrastive learning mitigates the influence of hyperparameters by maximizing feature distinctiveness. The FPH strategy prevents gradient updates in the projection network, enabling the focused training of the backbone networks. The proposed method shows a mean average precision (mAP) increase of 1.17 and 0.14 in 3D cross-modal object retrieval experiments using ModelNet10 and ModelNet40 datasets compared to state-of-the-art (SOTA) methods.
- 10 November 2024
- Published by MDPI AG in Applied Sciences
- Vol. 14 (22) , 10322
Open AccessSupervised Contrastive Learning for 3D Cross-Modal Retrieval
Interoperability between different virtual platforms requires the ability to search and transfer digital assets across platforms. Digital assets in virtual platforms are represented in different forms or modalities, such as images, meshes, and point clouds. The cross-modal retrieval of three-dimensional (3D) object representations is challenging due to data representation diversity, making common feature space discovery difficult. Recent studies have been focused on obtaining feature consistency within the same classes and modalities using cross-modal center loss. However, center features are sensitive to hyperparameter variations, making cross-modal center loss susceptible to performance degradation. This paper proposes a new 3D cross-modal retrieval method that uses cross-modal supervised contrastive learning (CSupCon) and the fixed projection head (FPH) strategy. Contrastive learning mitigates the influence of hyperparameters by maximizing feature distinctiveness. The FPH strategy prevents gradient updates in the projection network, enabling the focused training of the backbone networks. The proposed method shows a mean average precision (mAP) increase of 1.17 and 0.14 in 3D cross-modal object retrieval experiments using ModelNet10 and ModelNet40 datasets compared to state-of-the-art (SOTA) methods.
- 1 November 2024
- Published by Elsevier in Pattern Recognition
- Vol. 159
A unified framework for unsupervised action learning via global-to-local motion transformer
- 1 November 2024
- Published by Elsevier in Pattern Recognition
- Vol. 159
A unified framework for unsupervised action learning via global-to-local motion transformer
- 16 June 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1705-1714
MoST: Motion Style Transformer Between Diverse Action Contents
While existing motion style transfer methods are effective between two motions with identical content, their performance significantly diminishes when transferring style between motions with different contents. This challenge lies in the lack of clear separation between content and style of a motion. To tackle this challenge, we propose a novel motion style transformer that effectively disentangles style from content and generates a plausible motion with transferred style from a source motion. Our distinctive approach to achieving the goal of disentanglement is twofold: (1) a new architecture for motion style transformer with 'part-attentive style modulator across body parts' and ‘Siamese encoders that encode style and content features separately’; (2) style disentanglement loss. Our method outperforms existing methods and demonstrates exceptionally high quality, particularly in motion pairs with different contents, without the need for heuristic post-processing. Codes are available at https://github.com/Boeun-Kim/MoST.
- 16 June 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1705-1714
MoST: Motion Style Transformer Between Diverse Action Contents
While existing motion style transfer methods are effective between two motions with identical content, their performance significantly diminishes when transferring style between motions with different contents. This challenge lies in the lack of clear separation between content and style of a motion. To tackle this challenge, we propose a novel motion style transformer that effectively disentangles style from content and generates a plausible motion with transferred style from a source motion. Our distinctive approach to achieving the goal of disentanglement is twofold: (1) a new architecture for motion style transformer with 'part-attentive style modulator across body parts' and ‘Siamese encoders that encode style and content features separately’; (2) style disentanglement loss. Our method outperforms existing methods and demonstrates exceptionally high quality, particularly in motion pairs with different contents, without the need for heuristic post-processing. Codes are available at https://github.com/Boeun-Kim/MoST.
- 31 March 2024
- Published by Korean Society for Internet Information (KSII) in KSII Transactions on Internet and Information Systems
- Vol. 18 (3) , 670-684
3D Cross-Modal Retrieval Using Noisy Center Loss and SimSiam for Small Batch Training
KSII Transactions on Internet and Information Systems TIIS Paper Details
- 31 March 2024
- Published by Korean Society for Internet Information (KSII) in KSII Transactions on Internet and Information Systems
- Vol. 18 (3) , 670-684
3D Cross-Modal Retrieval Using Noisy Center Loss and SimSiam for Small Batch Training
KSII Transactions on Internet and Information Systems TIIS Paper Details
- 8 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 705-709
Omission-Free Inpainting: A Three-Stage Approach to Ensure Object Generation
This paper proposes a novel inpainting framework, omission-free inpainting, which ensures generating the desired object in the masked region. Despite recent advancements in text-driven and class-conditional inpainting models, they often fail to restore the missing object. To address this issue, the proposed framework includes a separate object generation stage, resulting in omission-free inpainting. The framework consists of three stages: background generation, object generation and refinement. The background generation stage restores a harmonious background with the surrounding pixels, while the object generation stage creates the desired object using a blending mask that allows the object to be influenced by the background’s color and brightness. Finally, the refinement stage blends the object and background to produce a visually realistic image. We compare the results qualitatively with the state-of-the-art methods, and our method outperforms the existing methods in CLIP score.
- 8 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 705-709
Omission-Free Inpainting: A Three-Stage Approach to Ensure Object Generation
This paper proposes a novel inpainting framework, omission-free inpainting, which ensures generating the desired object in the masked region. Despite recent advancements in text-driven and class-conditional inpainting models, they often fail to restore the missing object. To address this issue, the proposed framework includes a separate object generation stage, resulting in omission-free inpainting. The framework consists of three stages: background generation, object generation and refinement. The background generation stage restores a harmonious background with the surrounding pixels, while the object generation stage creates the desired object using a blending mask that allows the object to be influenced by the background’s color and brightness. Finally, the refinement stage blends the object and background to produce a visually realistic image. We compare the results qualitatively with the state-of-the-art methods, and our method outperforms the existing methods in CLIP score.
- 28 October 2022
- Published by Springer Nature
- p. 209-225
Global-Local Motion Transformer for Unsupervised Skeleton-Based Action Learning
- 28 October 2022
- Published by Springer Nature
- p. 209-225
Global-Local Motion Transformer for Unsupervised Skeleton-Based Action Learning