DWDayan Wu

Dayan Wu

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Dayan Wu has a total of 64 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    SEEDCAP: SEMANTIC EXPANSION AND ENTITY-DRIVEN ZERO-SHOT IMAGE CAPTIONING

    by Binbin Li, Shupei Xiao, Dayan Wu, Gengqi Yang, Siyu Jia, Zisen Qi

  • TrustSearch: Toward Secure and Efficient Reverse Image Search via SGX

    by Fang Zou, Jia Zhang, Jingwei Li, Dayan Wu, Xiong Li, Hongwei Li, Ting Chen, Xiaosong Zhang

    Outsourcing image management to a cloud should not only protect the confidentiality of image data, but also maintain the capability of reverse image search, which requires identifying the existing stored images that are similar to an input image. Previous studies build on cryptographic approaches to realize reverse image search on encrypted images, yet failing to achieve either security or performance. This paper explores trusted image search, which uses Intel SGX to realize reverse image search in an enclave, in order to provide security guarantees via SGX while performing search on plain data (inside the enclave) for performance. However, due to the resource limits of SGX, directly realizing the search process in the enclave incurs high performance overhead. We present TRUSTSEARCH, which implements various design approaches to mitigate the resource overhead of SGX. We evaluate TRUSTSEARCH using real-world image datasets, and show that it outperforms state-of-the-art approaches for search performance while preserving space efficiency for the enclave.

  • DualECot: Toward Explainable Emotion Causality Analysis in Conversations

    by Yulin Zhao, Dayan Wu, Zheng Lin, Weiping Wang

    Emotion Causality Analysis in Conversations is a critical research field that provides deep insights into the psychological and behavioral dynamics of individuals during interactions. Previous research has typically framed emotion causality analysis in conversations as an utterance-level classification problem. However, this coarse-grained approach often results in models capturing only superficial associations between emotional utterances and their cause utterances, rather than understanding the genuine causal relationships. This limitation undermines the generalizability of these models. However, this coarse-grained methodology often leads to models merely capturing the superficial correlations between emotional expressions and their corresponding cause statements, instead of delving into and comprehending the authentic causal connections. This limitation undermines the generalizability of these models. Additionally, existing methods are often criticized for their lack of interpretability, which significantly restricts their practical applications. In this paper, we introduce DualECot, a novel dual-view reasoning framework that leverages Chain-of-Thought to comprehensively analyze the emotional dynamics in conversations. In this paper, we present DualECot, a novel dual-view reasoning framework. By harnessing Chain- of- Thought, it enables a comprehensive analysis of emotional dynamics in conversations. Specifically, the framework consists of two complementary reasoning chains. The first, Diving into One's Heart, pinpoints potential emotion triggers and examines the target speaker's psychological state to uncover the process behind emotion evocation. The second, Stepping into Others' Shoes, investigates the emotional ripple effects of the target utterance on subsequent dialogue interactions. To tackle the shortage of interpretable reasoning data, we annotated reasoning chains on two benchmark datasets. We then assessed the framework's effectiveness via in-context learning and supervised experiments. Experimental results demonstrate that DualECot not only achieves superior performance, but also enhances explainability and generalization capabilities. We will release our annotation as an open resource for the community.Specifically, the framework consists of two complementary reasoning chains. The first, Diving into One's Heart, pinpoints potential emotion triggers and examines the target speaker's psychological state to uncover the process behind emotion evocation. The second, Stepping into Others' Shoes, investigates the emotional ripple effects of the target utterance on subsequent dialogue interactions. To tackle the shortage of interpretable reasoning data, we annotated reasoning chains on two benchmark datasets. We then assessed the framework's effectiveness via in-context learning and supervised experiments. Experimental results demonstrate that DualECot not only achieves superior performance, but also enhances explainability and generalization capabilities. We will release our annotation as an open resource for the community.

  • Enhancing facial privacy protection in customized diffusion models via masked attention erasure

    by Yisu Liu, Lin Wang, Wanqian Zhang, Jinyang An, Huashan Chen, Dayan Wu, Zheng Lin, Weiping Wang


  • Open Access

    Mitigating the Evolving Semantic Entanglement in Continual Learning of Vision-Language Models

    by Yiliang Zhu, Dayan Wu, Qinghang Su, Zexian Yang, Zheng Lin, Weiping Wang

  • Planning forward: Deep incremental hashing by gradually defrosting bits

    by Qinghang Su, Dayan Wu, Chenming Wu, Bo Li, Weiping Wang


  • Open Access

    Endogenous Recovery via Within-modality Prototypes for Incomplete Multimodal Hashing

    by Sa Zhu, Dayan Wu, Chenming Wu, Pengwen Dai, Bo Li

    Multimodal hashing projects multimodal data into compact binary codes, enabling rapid and storage-efficient retrieval of large-scale multimedia content. In practical scenarios, the issue of missing modality frequently arises when dealing with multimodal data. Existing incomplete multimodal hashing techniques directly recover missing modalities by neural networks, resulting in a disjointed representation space between the recovered and true data. In this paper, we present a novel recovery paradigm, namely Prototype-based Modality Completion Hashing (PMCH). Instead of directly synthesizing it from available modalities, PMCH adaptively aggregates associated within-modality prototypes to recover missing modality data. Specifically, PMCH introduces an within-modality prototype learning module to optimize representative prototypes for each modality. These prototypes act as recovery anchors and reside within the same representation space as their corresponding modality data. Subsequently, PMCH adaptively aggregates the associated within-modality prototypes with coefficients derived from the modality-specific Weight-Net. By utilizing prototypes from the same modality, the semantic disparity between the reconstructed and authentic data can be substantially diminished. Extensive experiments on three widely used benchmark datasets demonstrate that PMCH can effectively recover the missing modality, and attain state-of-the-art performance in both complete and incomplete multimodal retrieval scenarios. Code is available at https://github.com/Sasa77777779/PMCH.git.


  • HRACap: Lightweight Image Captioning via Hierarchical Retrieval-Augmented Prompt

    by Shupei Xiao, Binbin Li, Jia Siyu, Yang Gengqi, Qi Zisen, Wu Dayan, Wang Haiping, Ding Yu


  • TextSafety: Visual Text Vanishing via Hierarchical Context-Aware Interaction Reconstruction

    by Pengwen Dai, Jingyu Li, Dayan Wu, Peijia Zheng, Xiaochun Cao

    Privacy information existing in the scene text will be leaked with the spread of images in cyberspace. Vanishing the scene text from the image is a simple yet effective method to prevent privacy disclosure to the machine and the human. Previous visual text vanishing methods have achieved promising results but the performance still fell short of expectations for complicated-shape scene texts with various scales. In this paper, we propose a novel hierarchical context-aware interaction reconstruction method to make the visual text vanish in the natural scene image. To avoid the interference of the non-text regions, we narrow down the reconstruction regions by the guidance of the hierarchical refined text region masks, helping provide accurate position information. Meanwhile, we propose to learn the long-range context-aware interaction in a lightweight way, which can ensure the smoothing of the artifacts that are easily generated by the convolutional layers. To be more specific, we first simultaneously generate the coarse text region mask and the initially vanishing scene text image. Then, we obtain more accurate refined masks to better capture the locations of complicated-shape texts via a hierarchical mask generation network. Next, based on the refined masks, we exploit a channel-wise context-aware interaction mechanism to model the long-range relationships between the reconstruction region and the backgrounds for better removing the artifacts. Finally, we fuse the reconstructed text regions with the non-masked regions to obtain the ultimate protected image. Experiments on two frequently-used benchmarks SCUT-EnsText and SCUT-Syn demonstrate that our proposed method outperforms previous related methods by a large margin.