Faculty Profiles - Deguchi Daisuke

Papers - Deguchi Daisuke

Division display 1 - 20 of about 259 ／ All the affair displays >>

Semantic-Centric Alignment for Zero-shot Panoptic Segmentation with Limited Data Reviewed Open Access

Jialei Chen, Daisuke Deguchi, Dongyue Li, Xu Zheng, Seigo Ito, Hiroshi Murase, Qi Fan

International Journal of Computer Vision 2026.2

Pedestrians' Gaze Object Estimation in Traffic Scene Considering Relationships Among Multiple Pedestrians Reviewed

Hiroyuki Tatsumi, Daisuke Deguchi, Seigo Ito, Hiroshi Murase

Proceedings of International Workshop on Advanced Image Technology (IWAIT) 2026 2026.1

Split Matching for Inductive Zero-shot Semantic Segmentation Reviewed

Jialei Chen, Xu Zheng, Dongyue Li, CHONG YI, Seigo Ito, Danda Pani Paudel, Luc Van Gool, Hiroshi Murase, Daisuke Deguchi

Proceedings of British Machine Vision Conference (BMVC 2025) 2025.12

Semantic-driven Distillation for Semantic Segmentation with Unknown Classes

Jialei Chen, Dongyue Li, CHONG YI, Xu Zheng, Seigo Ito, Hiroshi Murase, Daisuke Deguchi

Proceedings of 2025 IEEE Conference on Intelligent Transportation Systems (ITSC2025) 2025.11

CQVPR: Landmark-aware Contextual Queries for Visual Place Recognition Reviewed

Dongyue Li, Jialei Chen, Seigo Ito, Hiroshi Murase, Daisuke Deguchi

Proceedings of 2025 IEEE Conference on Intelligent Transportation Systems (ITSC2025) 2025.10

CLIP-to-Seg Distillation for Zero-shot Semantic Segmentation Reviewed Open Access

Jialei Chen, Zhenzhen Quan, Chenkai Zhang, Xu Zheng, Daisuke Deguchi, Hiroshi Murase

IEEE Transactions on Circuits and Systems for Video Technology (TCSVT) 2025.9

Language：English Publishing type：Research paper (scientific journal)

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.

DOI： 10.1109/TCSVT.2025.3616588

Open Access

Semantic matters: A constrained approach for zero-shot video action recognition Reviewed

Zhenzhen Quan, Jialei Chen, Daisuke Deguchi, Jie Sun, Chenkai Zhang, Yujun Li, Hiroshi Murase

Pattern Recognition Vol. 162 page： 111402 2025.6

A Cross-Modal Knowledge Distillation Approach for RGB-to-Infrared Video Action Recognition Reviewed

Zhenzhen Quan, Daisuke Deguchi, Jialei Chen, Chenkai Zhang, Yujun Li, Seigo Ito, Hiroshi Murase

Proceedings of Korea-Japan Joint Workshop on Frontiers of Computer Vision (FCV2025) 2025.2

Multi-group Vision Semantic Centroid for Semantic Segmentation Reviewed

Jialei Chen, Daisuke Deguchi, Chenkai Zhang, Zhenzhen Quan, Seigo Ito, Hiroshi Murase

Proceedings of Korea-Japan Joint Workshop on Frontiers of Computer Vision (FCV2025) 2025.2

Language：English Publishing type：Research paper (international conference proceedings)

Multi-media processing has achieved great success based on
semantic segmentation. Semantic segmentation can be viewed as pixel-
clustering based on semantic prototypes. However, existing methods fo-
cus more on consistent semantics while ignoring the consistency in vi-
sion, making this task challenging. Motivated by the success of discrete
visual representation learning, we propose Multi-group Visual Semantic
Centroid (MVSC) to better cluster the pixels while maintaining consis-
tent semantics of the dense features for any image encoder. Specifically,
we randomly initialize multiple groups of prototypes as multi-groups in
visual space. The visual features are also randomly split into the same
groups and forced to be aligned with the corresponding prototypes. Then
these visual prototypes are projected into the semantic space and super-
vised by the same classifier as the dense features. Compared with existing
methods, MVSC further considers the visual space and thus facilitates
the task. Experimental results on COCO-Stu! show great improvements
compared with previous methods.

CROCODILE: Crop-based Contrastive Discriminative Learning for Enhancing Explainability of End-to-End Driving Models Reviewed

Chenkai Zhang, Daisuke Deguchi, Jialei Chen, Zhenzhen Quan, Hiroshi Murase

Proceedings of ACCV Workshop on Machine Learning and Computing for Visual Semantic Analysis (MLCSA2024) page： 378 - 393 2024.12

Early prediction of At-risk students through Learning-Activity Forecasting Reviewed

Yuya Ozaki, Daisuke Deguchi, Haruya Kyutoku, Hiroshi Murase

Proceedings of 32nd International Conference on Computers in Education (ICCE2024) 2024.11

Toward Explainable End-to-End Driving Models via Simplified Objectification Constraints Reviewed Open Access

Chenkai Zhang, Daisuke Deguchi, Jialei Chen, Hiroshi Murase

IEEE Transactions on Intelligent Transportation Systems Vol. 25 ( 10 ) page： 14521 - 14534 2024.10

Language：English Publishing type：Research paper (scientific journal)

The end-to-end driving models (E2EDMs) convert environmental information into driving actions using a complex transformation which makes E2EDMs have high prediction accuracy. Due to the black-box nature of transformation, the E2EDMs have low explainability. To solve this problem, explanation methods are used to generate explanations for observation. Based on current explanation methods, previous studies tried to further improve the explainability of E2EDMs by integrating an object detection module, however, these methods have many problems: Firstly, due to the requirement of the object detection module, they lack flexibility. Secondly, they neglect an essential property, i.e. , simplicity, to improve explainability. In this paper, since humans prefer object-level and simple explanations in driving tasks, we argue that explainability is decided by two properties which are the objectification degree (the extent to which driving related-object features are utilized) and simplification degree (the simplicity of the explanation), thus we propose Simplified Objectification Branches (SOB) to improve the explainability of E2EDMs. Firstly, this structure could be integrated into any existing E2EDMs and thus have high flexibility. Secondly, the SOB explicitly improves the simplification degree without sacrificing the objectification degree of the explanations. By designing several indicators, i.e. , heatmap satisfaction, driving action reproduction score, deception level, etc. , we proved that SOB could help E2EDMs generate better explanations. Notably, the SOB could also further enhance E2EDMs’ prediction accuracy.

DOI： 10.1109/TITS.2024.3385754

Open Access

Frozen is better than learning: A new design of prototype-based classifier for semantic segmentation Reviewed Open Access

Jialei Chen, Daisuke Deguchi, Chenkai Zhang, Xu Zheng, Hiroshi Murase

Pattern Recognition Vol. 152 2024.8

Language：English Publishing type：Research paper (scientific journal)

Semantic segmentation models comprise an encoder to extract features and a classifier for prediction. However, the learning of the classifier suffers from the ambiguity which is caused by two factors: (1) the weights of a classifier for similar categories may have positive similarities lowing the performance for similar categories, named correlation ambiguity, and (2) the classifier is prone to predict the category with a larger L2 norm and vice versa, termed prior ambiguity. To comedy the issues, we propose Category-Basis Prototype (CBP), frozen and mutually orthogonalized prototypes with equal L2 norm. Orthogonalization prevents the prototypes from being similar to each other and the equality decouples the prediction from the L2 norm. To better shape the feature space, we propose Online Centroid Contrastive Loss (OCCL) equipped with centroid and category- level losses. Experiments show that our method yields compelling results over two widely applied benchmarks indicating the effectiveness of our methods.

DOI： 10.1016/j.patcog.2024.110431

Open Access

Eye-Contact Transformer: Eye-Contact Detection of Distant Pedestrians Considering Scene Context

Ryusei Hata, Daisuke Deguchi, Takatsugu Hirayama, Yasutomo Kawanishi, Hiroshi Murase

IEICE Transactions on Information and Systems Vol. J107-D ( 4 ) page： 231 - 242 2024.4

Centroid Module for Shaping Feature Space in Semantic Segmentation

Jialei Chen, Daisuke Deguchi, Chenkai Zhang, Hiroshi Murase

Proceedings of 3rd Asia Conference on Cloud Computing, Computer Vision and Image Processing (3CVIP 2024) 2024.3

Subjective Baggage-Weight Estimation Based on Human Walking Behavior Reviewed Open Access

Masaya Mizuno, Tomohiro Fujita, Yasutomo Kawanishi, Daisuke Deguchi, Hiroshi Murase

IEEE Access Vol. 12 page： 39390 - 39398 2024.3

Category-Level Object Pose Estimation in Heavily Cluttered Scenes by Generalized Two-Stage Shape Reconstructor Reviewed Open Access

Hiroki Tatemichi, Yasutomo Kawanishi, Daisuke Deguchi, Ichiro Ide, Hiroshi Murase

IEEE Access Vol. 12 page： 33440 - 33448 2024.3

Comprehensive Evaluation of End-to-End Driving Model Explanations for Autonomous Vehicles Reviewed

Chenkai Zhang, Daisuke Deguchi, Jialei Chen, Hiroshi Murase

Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISAPP) 2024 Vol. 2 ( 2184-4321 ) page： 509 - 518 2024.2

Language：English Publishing type：Research paper (international conference proceedings)

Deep learning technology has rapidly advanced, leading to the development of End-to-End driving models (E2EDMs) for autonomous vehicles with high prediction accuracy. To comprehend the prediction results of these E2EDMs, one of the most representative explanation methods is attribution-based. There are two kinds of attribution-based explanation methods: pixel-level and object-level. Usually, the heatmaps illustrate the importance of pixels and objects in the prediction results, serving as explanations for E2EDMs. Since there are many attribution-based explanation methods, evaluation methods are proposed to determine which one is better at improving the explainability of E2EDMs.
Fidelity measures the explanation's faithfulness to the model's prediction method, which is a bottommost property. However, no evaluation method could measure the fidelity difference between object-level and pixel-level explanations, making the current evaluation incomplete. In addition, without considering fidelity, previous evaluation methods may advertise manipulative explanations that solely seek human satisfaction (persuasibility). Therefore, we propose an evaluation method that further considers fidelity, our method enables a comprehensive evaluation that proves the object-level explanations genuinely outperform pixel-level explanations in fidelity and persuasibility, thus could better improve the explainability of the E2EDMs.

Pedestrian's Gaze Object Detection in Traffic Scene Reviewed

Hiroto Murakami, Jialei Chen, Daisuke Deguchi, Takatsugu Hirayama, Yasutomo Kawanishi, Hiroshi Murase

Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISAPP) 2024 Vol. 2 page： 333 - 340 2024.2

Language：English Publishing type：Research paper (international conference proceedings)

In this paper, we present a new task of detecting an object that a target pedestrian is gazing at in a traffic scene called PEdestrian's Gaze Object (PEGO). We argue that the detection of gaze object can provide important information for pedestrian's behavior prediction and can contribute to the realization of automated vehicles. For this task, we construct a dataset of in-vehicle camera images with annotations of the objects that pedestrians are gazing at. Also, we propose a Transformer-based method called PEGO Transformer to solve the PEGO detection task. The PEGO Transformer directly performs gaze object detection with the utilization of whole-body features without a high-resolution head image and a gaze heatmap which the traditional methods rely on. Experimental results showed that the proposed method could estimate pedestrian's gaze object accurately even if various objects exist in the scene.

Texture-guided transfer learning for low-quality face recognition Reviewed Open Access

Meng Zhang, Rujie Liu, Daisuke Deguchi, Hiroshi Murase

IEEE Transactions on Image Processing 2023.11

Language：English Publishing type：Research paper (scientific journal)

Although many advanced works have achieved significant progress for face recognition with deep learning and large-scale face datasets, low-quality face recognition remains a challenging problem in real-word applications, especially for unconstrained surveillance scenes. We propose a texture-guided (TG) transfer learning approach under the knowledge distillation scheme to improve low-quality face recognition performance. Unlike existing methods in which distillation loss is built on forward propagation; e.g., the output logits and intermediate features, in this study, the backward propagation gradient texture is used. More specifically, the gradient texture of low-quality images is forced to be aligned to that of its high-quality counterpart to reduce the feature discrepancy between the high-and low-quality images. Moreover, attention is introduced to derive a soft-attention (SA) version of transfer learning, termed as SA-TG, to focus on informative regions. Experiments on the benchmark low-quality face DB’s TinyFace and QMUL-SurFace confirmed the superiority of the proposed method, especially more than 6.3% Rank1 accuracy improvement is achieved on TinyFace.

DOI： 10.1109/TIP.2023.3335830

Open Access

　 PREV - NEXT 　

<Deguchi Daisuke>

Papers - Deguchi Daisuke