Daisuke Deguchi
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Daisuke Deguchi has a total of 71 co-authors. Here are the top co-authors by number of shared publications:
- Hiroshi Murase
Nagoya University, Nagoya, Japan
20 shared publications - Yasutomo Kawanishi
Guardian Robot Project R-IH, RIKEN
11 shared publications - Ichiro Ide
Nagoya University, Nagoya, Japan
10 shared publications - Jialei Chen
Graduate School of Informatics, Nagoya University, Nagoya, Japan
10 shared publications - Takatsugu Hirayama
Nagoya University
8 shared publications - Chenkai Zhang
Graduate School of Informatics, Nagoya University, Nagoya, Japan
7 shared publications
Latest Publications
- Open Access
- 28 July 2026
- Published by Elsevier in SSRN Electronic Journal
Beyond Frame-level Representation for Zero-shot Video Action Recognition
Zero-shot video action recognition aims to recognize unseen action categories by transferring knowledge from seen categories through vision-language alignment. However, existing CLIP-based methods inherit image-centric representations, limiting their ability to capture hierarchical temporal dependencies and coherent action semantics across video sequences. To address this limitation, we propose HTM-CLIP, a unified vision-language framework that enhances frozen CLIP through hierarchical temporal modeling and conditional semantic enhancement. Specifically, hierarchical temporal modeling captures short-, mid-, and long-range dependencies in a progressive manner, enabling efficient video representation learning without global attention. Meanwhile, conditional semantic enhancement facilitates early visual-semantic interaction and category-aware feature refinement via gated prompt modulation. Together, these components learn temporally coherent and semantically discriminative video representations, improving vision-language alignment and zero-shot generalization. Extensive experiments demonstrate that HTM-CLIP consistently outperforms state-of-the-art methods while maintaining competitive computational efficiency.
- 28 July 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessBeyond Frame-level Representation for Zero-shot Video Action Recognition
Zero-shot video action recognition aims to recognize unseen action categories by transferring knowledge from seen categories through vision-language alignment. However, existing CLIP-based methods inherit image-centric representations, limiting their ability to capture hierarchical temporal dependencies and coherent action semantics across video sequences. To address this limitation, we propose HTM-CLIP, a unified vision-language framework that enhances frozen CLIP through hierarchical temporal modeling and conditional semantic enhancement. Specifically, hierarchical temporal modeling captures short-, mid-, and long-range dependencies in a progressive manner, enabling efficient video representation learning without global attention. Meanwhile, conditional semantic enhancement facilitates early visual-semantic interaction and category-aware feature refinement via gated prompt modulation. Together, these components learn temporally coherent and semantically discriminative video representations, improving vision-language alignment and zero-shot generalization. Extensive experiments demonstrate that HTM-CLIP consistently outperforms state-of-the-art methods while maintaining competitive computational efficiency.
- Open Access
- 25 July 2026
- Published by Elsevier in SSRN Electronic Journal
Not All Pixels Are Equal: Confidence-GuidedAttention for Feature Matching
Semi-dense feature matching methods have been significantly advanced by leveraging attention mechanisms to extract discriminative descriptors. However, most existing approaches treat all pixels equally during attention computations, which can potentially introduce noise and redundancy from irrelevant regions. To address this issue, we propose a Confidence-guided semi-dense Matcher (CgMatcher) that adaptively prunes attention weights for each pixel based on precomputed matching confidence maps. These maps are generated by evaluating the mutual similarity between feature pairs extracted from the backbone, where high confidence indicates a high potential for matching. Then the attention is refined through two steps: (1) a confidence-guided bias is introduced to adaptively adjust the attention distributions for each query pixel, avoiding irrelevant interactions between non-overlap pixels; (2) the corresponding confidence map is additionally employed to rescale value features during feature aggregation, attenuating the influence of uncertain regions. Moreover, a classification loss is introduced to encourage the backbone's features to discriminate between matchable and non-matchable regions. Extensive experiments on three benchmarks demonstrate that the proposal outperforms existing state-of-the-art methods.
- 25 July 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessNot All Pixels Are Equal: Confidence-GuidedAttention for Feature Matching
Semi-dense feature matching methods have been significantly advanced by leveraging attention mechanisms to extract discriminative descriptors. However, most existing approaches treat all pixels equally during attention computations, which can potentially introduce noise and redundancy from irrelevant regions. To address this issue, we propose a Confidence-guided semi-dense Matcher (CgMatcher) that adaptively prunes attention weights for each pixel based on precomputed matching confidence maps. These maps are generated by evaluating the mutual similarity between feature pairs extracted from the backbone, where high confidence indicates a high potential for matching. Then the attention is refined through two steps: (1) a confidence-guided bias is introduced to adaptively adjust the attention distributions for each query pixel, avoiding irrelevant interactions between non-overlap pixels; (2) the corresponding confidence map is additionally employed to rescale value features during feature aggregation, attenuating the influence of uncertain regions. Moreover, a classification loss is introduced to encourage the backbone's features to discriminate between matchable and non-matchable regions. Extensive experiments on three benchmarks demonstrate that the proposal outperforms existing state-of-the-art methods.
- Open Access
- 20 July 2026
- Published by Springer Nature in International Journal of Computer Vision
- Vol. 134 (8) , 353
Correction: Training-Free Open-Vocabulary Semantic Segmentation with Context Pyramid Refinement
- 20 July 2026
- Published by Springer Nature in International Journal of Computer Vision
- Vol. 134 (8) , 353
Open AccessCorrection: Training-Free Open-Vocabulary Semantic Segmentation with Context Pyramid Refinement
- 23 May 2026
- Published by Springer Nature
- p. 16-29
Multi-Group Vision Semantic Centroid for Semantic Segmentation
- 23 May 2026
- Published by Springer Nature
- p. 16-29
Multi-Group Vision Semantic Centroid for Semantic Segmentation
- 23 May 2026
- Published by Springer Nature
- p. 30-42
A Cross-Modal Knowledge Distillation Approach for RGB-to-Infrared Video Action Recognition
- 23 May 2026
- Published by Springer Nature
- p. 30-42
A Cross-Modal Knowledge Distillation Approach for RGB-to-Infrared Video Action Recognition
- 21 May 2026
- Published by Springer Nature in International Journal of Computer Vision
- Vol. 134 (6) , 281
Training-Free Open-Vocabulary Semantic Segmentation with Context Pyramid Refinement
- 21 May 2026
- Published by Springer Nature in International Journal of Computer Vision
- Vol. 134 (6) , 281
Training-Free Open-Vocabulary Semantic Segmentation with Context Pyramid Refinement
- 16 February 2026
- Published by Springer Nature in International Journal of Computer Vision
- Vol. 134 (3) , 122
Semantic-Centric Alignment for Zero-shot Panoptic Segmentation with Limited Data
- 16 February 2026
- Published by Springer Nature in International Journal of Computer Vision
- Vol. 134 (3) , 122
Semantic-Centric Alignment for Zero-shot Panoptic Segmentation with Limited Data
- Open Access
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 28, 1-12
BiXFormer: A Robust Framework for Maximizing Modality Effectiveness in Multi-Modal Semantic Segmentation
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 28, 1-12
Open AccessBiXFormer: A Robust Framework for Maximizing Modality Effectiveness in Multi-Modal Semantic Segmentation
- Open Access
- 1 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. 36 (3) , 3665-3678
CLIP-to-Seg Distillation for Zero-Shot Semantic Segmentation
CLIP has greatly advanced zero-shot segmentation by leveraging its strong visual-language association and generalization capability. However, directly adapting CLIP for segmentation often yields suboptimal results due to inconsistencies between image and pixel-level prediction objectives. Additionally, merely combining segmentation and CLIP models often leads to disjoint optimization, introducing significant computational overhead and additional parameters. To address these issues, we propose a novel CLIP-to-Seg Distillation approach, incorporating global and local distillation to flexibly transfer CLIP’s powerful zero-shot generalization capability to existing closed-set segmentation models. Global distillation leverages CLS tokens to condense segmentation features and distills high-level concepts to the segmentation model via image-level features. Local distillation adapts CLIP’s local semantic transferability to dense prediction tasks using object-level features, aided by pseudo-mask generation for latent class mining. To further generalize the CLIP-distilled segmentation model, we generate latent text embeddings for the mined latent classes by coordinating their text embeddings and dense features. Our method equips existing closed-set segmentation models with strong generalization capabilities for open concepts through effective and flexible CLIP-to-Seg distillation. Without relying on the CLIP model or introducing extra inference overhead, our method seamlessly integrates into existing closed-set segmentation models and enables zero-shot capability, achieving state-of-the-art performance on multiple benchmarks.
- 1 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. 36 (3) , 3665-3678
Open AccessCLIP-to-Seg Distillation for Zero-Shot Semantic Segmentation
CLIP has greatly advanced zero-shot segmentation by leveraging its strong visual-language association and generalization capability. However, directly adapting CLIP for segmentation often yields suboptimal results due to inconsistencies between image and pixel-level prediction objectives. Additionally, merely combining segmentation and CLIP models often leads to disjoint optimization, introducing significant computational overhead and additional parameters. To address these issues, we propose a novel CLIP-to-Seg Distillation approach, incorporating global and local distillation to flexibly transfer CLIP’s powerful zero-shot generalization capability to existing closed-set segmentation models. Global distillation leverages CLS tokens to condense segmentation features and distills high-level concepts to the segmentation model via image-level features. Local distillation adapts CLIP’s local semantic transferability to dense prediction tasks using object-level features, aided by pseudo-mask generation for latent class mining. To further generalize the CLIP-distilled segmentation model, we generate latent text embeddings for the mined latent classes by coordinating their text embeddings and dense features. Our method equips existing closed-set segmentation models with strong generalization capabilities for open concepts through effective and flexible CLIP-to-Seg distillation. Without relying on the CLIP model or introducing extra inference overhead, our method seamlessly integrates into existing closed-set segmentation models and enables zero-shot capability, achieving state-of-the-art performance on multiple benchmarks.
- 18 June 2025
- Published by Springer Nature
- p. 300-311
Single-agent vs. Multi-agent LLM Strategies for Automated Student Reflection Assessment
- 18 June 2025
- Published by Springer Nature
- p. 300-311
Single-agent vs. Multi-agent LLM Strategies for Automated Student Reflection Assessment