Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileAcademic lineage
Possible advisorsa guess from early papers, not confirmed
- Masaru KitsuregawaSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works91 from public data
- CosFace: Large Margin Cosine Loss for Deep Face Recognition3,113
This paper reformulates the softmax loss as a cosine loss by L2 normalizing both features and weight vectors to remove radial variations, based on which acosine margin term is introduced to further maximize the decision margin in the angular space, and achieves minimum intra-class variance and maximum inter- class variance by virtue of normalization and cosine decision margin maximization.
- DiffIR: Efficient Diffusion Model for Image Restoration476
This work proposes an efficient DM for IR (DiffIR), which consists of a compact IR prior extraction network (CPEN), dynamic IR transformer (DIRformer), and denoising network, which can adopt a joint optimization of CPENS2, DIRformer, and denoising network, which can further reduce the estimation error influence.
- FreeSeg: Unified, Universal and Open-Vocabulary Image Segmentation156
FreeSeg is proposed, a generic framework to accomplish Unified, Universal and Open-Vocabulary Image Segmentation, which optimizes an all-in-one network via one-shot training and employs the same architecture and parameters to handle diverse segmentation tasks seamlessly in the inference procedure.
- Orthogonal Deep Features Decomposition for Age-Invariant Face Recognition130
A novel approach (namely, Orthogonal Embedding CNNs, or OE-CNNs) is proposed to learn the age-invariant deep face features to reduce the intra-class discrepancy caused by the aging.
- Mobile Edge Computing, Metaverse, 6G Wireless Communications, Artificial Intelligence, and Blockchain: Survey and Their Convergence82
This paper investigates the computational paradigms used to meet the stringent requirements of modern applications of mobile edge computing, and presents the motivation of MEC-based Metaverse and introduces the applications of M EC to the Metaverse.
- 82
- LAPIN: Effective Sequential Pattern Mining Algorithms by Last Position Induction for Dense Databases82
This work proposes a series of novel algorithms, called the LAst Position INduction (LAPIN) sequential pattern mining, which is based on the simple idea that the last position of an item, α is the key to judging whether or not a frequent k-length sequential pattern can be extended to be a frequent (k+1)-length pattern by appending the item α to it.
- Link Based Clustering of Web Search Results81
A new approach to cluster search results returned from Web search engine using link analysis based on common links shared by pages using co-citation and coupling analysis, which is based on document clustering algorithms in IR.
- Evaluating contents-link coupled web page clustering for web search results79
A contents-link coupled clustering algorithm that clusters web pages by combining contents and link analysis is proposed that is effective and promising and applies to cluster web search results.
- Open-Vocabulary Segmentation with Semantic-Assisted Calibration78
This paper studies open-vocabulary segmentation through calibrating in-vocabulary and domain-biased embedding space with generalized contextual prior of CLIP and presents a Semantic-assisted CAlibration Network (SCAN), which achieves state-of-the-art performance on all popular open-vocabulary segmentation benchmarks.
- AniGAN: Style-Guided Generative Adversarial Networks for Unsupervised Anime Face Generation75
A novel GAN-based anime-face translator, called AniGAN, is proposed to synthesize high-quality anime-faces which are style-consistent with a given reference anime- Face, while preserving the global structure of the source photo-face.
- CreatiLayout: Siamese Multimodal Diffusion Transformer for Creative Layout-to-Image Generation74
The Layout Designer is introduced, which taps into the potential of large language models in layout planning, transforming them into experts in layout generation and optimization and form CreatiLayout-a systematic solution that integrates the layout model, dataset, and planner for creative layout-to-image generation.
- Global Knowledge Calibration for Fast Open-Vocabulary Segmentation64
A core idea of preserving the generalizable representation when fine-tuning on known classes is proposed, and a text diversification strategy that generates a set of synonyms for each training category is introduced, which prevents the learned representation from collapsing onto specific known category names.
- Toward blockchain and fog computing collaborative design and manufacturing platform: Support customer view60
A blockchain-enabled fog computing-based collaborative design and manufacturing platform is proposed to develop triple communication and cooperation among the manufacturing section, design section, and customers in a trustable environment to improve the data integrity and security problem.
- 59
- A Survey of Mobile Edge Computing for the Metaverse: Architectures, Applications, and Challenges52
This paper proposes a comprehensive survey of the MEC-based Metaverse, and introduces the potential future directions for developing Metaverse systems, e.g., BoundlessXR and CloudXR.
- Human De-occlusion: Invisible Perception and Recovery for Humans48
A two-stage framework is proposed to estimate the invisible portions and recover the content inside of human deocclusion and a stacked network structure is devised to refine inaccurate masks from a general instance segmentation model and predict integrated masks simultaneously.
- 48
- Universal Segmentation at Arbitrary Granularity with Language Instruction47
Unified segmentation model UniLSeg is presented, a universal segmentation model that can perform segmentation at any semantic level with the guidance of language instructions, surpassing both specialist and unified segmentation models.
- 43
- 43
- Structured Sparsity Learning for Efficient Video Super-Resolution40
In SSL, a structured pruning scheme called Structured Sparsity Learning (SSL) is developed according to the properties of VSR, which can significantly outperform recent methods quantitatively and qualitatively.
- Basic Binary Convolution Unit for Binarized Image Restoration Network36
This study reconsider components in binary convolution, such as residual connection, BatchNorm, activation function, and structure, for IR tasks, and finds that residual connection can reduce the information loss caused by binarization, and BBCU can serve as a basic unit for binarized IR networks.
- DreamOmni: Unified Image Generation and Editing34
This work introduces DreamOmni, a unified model for image generation and editing that integrates both T2I models and various editing tasks, and develops a synthetic data pipeline using sticker-like elements to synthesize accurate, high-quality datasets efficiently.
- On Combining Link and Contents Information for Web Page Clustering33
The proposed approach automatically clusters the web search results into high quality, semantically meaningful groups in a concise, easy-to-interpret hierarchy with tagging terms and results show that the proposed approach is effective and promising.
- DiffI2I: Efficient Diffusion Model for Image-to-Image Translation32
Compared to traditional DMs, the compact IPR enables DiffI2I to obtain more accurate outcomes and employ a lighter denoising network and fewer iterations, and it is demonstrated that DiffI2I achieves SOTA performance while significantly reducing computational burdens.
- 29
- An Effective System for Mining Web Log26
This work proposes an efficient sequential mining algorithm (LAPIN_WEB: LAst Position INduction for WEB log), an extension of previous LAPIN algorithm to extract user access patterns from traversal path in Web logs, which outperforms well-known PrefixSpan by up to an order of magnitude on real Web log datasets.
- 18
- Automatic Animation of Hair Blowing in Still Portrait Photos17
This work innovatively treats hair wisp extraction as an instance segmentation problem, where a hair wisp is referred to as an instance, and proposes a wisp-aware animation module that animates hair wisps with pleasing motions without noticeable artifacts.
- 17
- Offloading and Quality Control for AI Generated Content Services in 6G Mobile Edge Computing Networks13
A joint optimization scheme for offloading decisions, computation time, and diffusion steps of the diffusion models in the reverse diffusion stage is presented and Experimental outcomes definitively show that the algorithm put forward outperforms baseline methods in terms of joint optimization performance.
- 13
- DreamVE: Unified Instruction-based Image and Video Editing10
This work introduces DreamVE, a unified model for instruction-based image and video editing, and uses generative model-based data to further fine-tune DreamVE and design an efficient and powerful editing framework for DreamVE.
- A Survey on Private Transformer Inference9
This paper reviews recent advancements in PTI, analyzing state-of-the-art solutions, their challenges, and potential improvements, and proposes evaluation guidelines to assess resource efficiency and privacy guarantees, aiming to bridge the gap between high-performance inference and data privacy.
- Effective Sequential Pattern Mining Algorithms for Dense Database9
A series of novel algorithms, called the LAst Position INduction (LAPIN) sequential pattern mining, which is based on the simple idea that the last position of an item, fi, is the key to judging whether or not a frequent k-length sequential pattern can be extended to be a frequent (k+1)-length pattern by appending the item fi to it.
- Private Transformer Inference in MLaaS: A Survey6
This paper reviews recent PTI advancements and introduces a structured taxonomy and evaluation framework for PTI, focusing on balancing resource efficiency with privacy and bridging the gap between high-performance inference and data privacy.
- Semantic-Assisted Object Clustering for Multi-Modal Referring Video Segmentation6
Semantic-assisted Object Cluster network (SOC) and the improved SOC++ are presented, which unifies temporally selective interaction and cross-modal alignment to achieve video-level understanding and outperforms state-of-the-art competitors on all benchmarks by a remarkable margin.
- Pixelwise Adaptive Discretization with Uncertainty Sampling for Depth Completion6
This paper is the first to consider the difference between pixels and propose Pixelwise Adaptive Discretization to generate the tailored depth hypotheses for each pixel, which allows the discrete depth hypotheses to be concentrated around the ground-truth of each pixel as much as possible.
- 6
- 6
- 6
- Enhancing Contents-Link Coupled Web Page Clustering and Its Evaluation6
An experimental study on enhancing coupling of links and contents analysis of Web pages for robust clustering and introduces two techniques: in-link reinforcement and anchor window analysis to improve the adaptability of contents-link coupled clustering.
- Laser Image Enhancement Algorithm Based on Improved EnlightenGAN5
An enhancement algorithm based on the EnlightenGAN network is proposed to improve the signal-to-noise ratio and contrast of laser images with richer image details and provides a new idea for pre-processing laser images.
- 5
- An efficient scheme for SDN state consistency verification in cloud computing environment5
The results show that the proposed scheme can detect security threats and label the abnormal forwarding path in real time to ensure the network state consistency, while introducing negligible performance overhead.
- Identifying Topic Experts and Topic Communities in the Blogspace5
The challenge of efficient identifying topic experts is discussed and a novel model is proposed to quantify topic experts and a new approach to identify the related blog communities on that topic is proposed.
- 4
- Application of multidimensional gait feature fusion algorithm in gait assessment for patients with knee osteoarthritis4
The multidimensional gait feature fusion algorithm effectively overcomes the limitation of assessing either coordination or complexity in isolation, providing an interpretable quantitative tool for analyzing KOA pathological mechanisms and dynamically monitoring post-TKA rehabilitation.
- 4
- 4
- Distill-AER: Fine-Grained Address Entity Recognition from Spoken Dialogue via Knowledge Distillation4
- Use noisy link analysis to improve web search4
This work proposes an approach and corresponding penalty strategies to both detect and handle "noisy link" effectively and automatically and compared with other related works to demonstrate that this approach is rather effective in noisy link filtering and could improve the final search results significantly.
- Effective Algorithms for Sequential Pattern Mining4
This work has proposed a series of novel algorithms, called the LAst Position INduction (LAPIN) sequential pattern mining, which is based on the simple idea that the last position of an item is the key to judging whether or not a frequent k-length sequential pattern can be extended to be a frequent (k+1)-length pattern by appending the item α to it.
- Infrastructure for Rapid Open Knowledge Network Development3
A National Science Foundation Convergence Accelerator project is described to build a set of Knowledge Network Programming Infrastructure systems to address the issue of frustratingly slow building, using, and scaling large knowledge networks.
- 3
- 3
- The Benefits of Virtual Reality Travel on Symptom Burden and Mental Health of Hemodialysis Patients2
A single VR travel session can alleviate symptom burden, enhance subjective well-being, and reduce perceived stress in HD patients, suggesting that this experience could be integrated into care routines as a non-pharmacological intervention to improve symptom burden and mental health in HD patients.
- 1st Place Solution for 5th LSVOS Challenge: Referring Video Object Segmentation2
This work first obtain binary mask sequences from the RVOS models, and proposes Two-Stage Multi-Model Fusion strategy, which achieves 75.7% J&F on Ref-Youtube-VOS validation set and 70% J&F on test set, which ranks 1st place on 5th Large-scale Video Object Segmentation Challenge (ICCV 2023) track 3.
- 2
- 2
- Research on Interactive Design of Interface Layout of Idle Resource Transactional Websites2
This paper analyzes and classifies the idle resources transactional website, and proposes the interface layout design evaluation index, which proves that interface layouts design is a very important factor for user experience.
- Combating Link Spam by Noisy Link Analysis2
The experimental results demonstrate that the proposed "noisy link" analysis is very effective on both spam page filtering and final ranking improvement, and compared with two other related works (TrustRank and Site-level Noise removal) on two real web datasets.
- C4-2: Combining Link and Contents in Clustering Web Search Results to Improve Information Interpretation2
By filtering some irrelevant pages, the proposed approach clusters high quality pages in web search results into semantically meaningful groups with additional tagging keywords to facilitate users accessing and understanding.
- 2
- 1
- Metaverse and Business Economics: Theoretical Foundations and Research Prospects1
The metaverse has broken through the limitations of time and space and has a certain industrial foundation in entertainment, education, and psychotherapy, and the unique economic value of the metauniverse is analyzed.
- 1
- 1
- 1
- A new PSO-ABC wireless sensor network coverage optimization algorithm1
By using the proposed PSO-ABC hybrid algorithm to search for the optimal deployment scheme of nodes, coverage blind areas and coverage redundancy in the network can be significantly avoided.
- 1
- Automatic Recognition and Segmentation of Composite User Stories Based on Semantic Analysis1
A method for identifying and segmenting compound stories based on semantic analysis, which includes analyzing the structural and semantic features of compound stories, building a feature dictionary, and automatically recognizing compound storiesbased on HanLP is proposed.
- 1
- –
- Design of a lightweight crop disease monitoring platform based on UAV hyperspectral imaging technology–
The platform successfully monitored diseases such as black mold, rust, and blight in the application of corn test fields, and timely and effectively prevented and controlled the spread of the diseases, which is of great significance to improve crop yield and quality.
- –
- Flow and Continuance Intention in Augmented Reality (AR) Shopping: A Systematic Literature Review–
It is suggested that retailers and developers should enhance enjoyment and concentration and optimize control and leverage curiosity-driven exploration to strengthen user retention and highlight theoretical and methodological gaps.
- –
- –
- Research on map building and positioning technology for AGV navigation system–
Theoretical analyses and experimental results show that the improved Cartographer algorithm improves the map building efficiency and map accuracy.
- –
- –
- Under the background of big data - investigation and research on the learning effect of college students' dual-track system–
It is believed that the online teaching platform should establish a rich knowledge database to provide students with some resources outside the classroom, which can mobilize students' interest in learning, and at the same time, it should have a more complete evaluation system, so as not to increase students' inertia.
- –
- Equivalences in Variables Refinement for Differential Semi-algebraic Hybrid Systems–
This paper proposed differential semi-algebraic event structure as the system model, and then built the variables refinement theory of differential Semi-Algebraic hybrid system based on the various equivalence relationship.
- –
- DSP: Deep Sign Prediction in Signed Social Networks–
This paper proposes a Deep Sign Prediction (DSP) method, which uses deep learning technology to capture the structure information of the signed social networks and shows that the proposed DSP method is very effective and outperforms other methods in terms of four metrics.
- –
- –
- –
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-14. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.