Zhiqiang Yuan

Location: Beijing, China
Work: Wechat AI, Tencent
Email: yuanzhiqiang19@mails.ucas.ac.cn
GoogleScholar Github

Graduated with a bachelor's degree from Harbin Engineering University in 2019, and obtained a doctorate in engineering from Aerospace Information Research Institute, Chinese Academy of Sciences in 2024. From 2024 to present, working at WeChat AI, long-term committed to the research and application of deep learning methods under sample and resource-constrained conditions.

Research


Deep Learning Under Limited Samples

  • Remote Sensing Cross-modal Retrieval Method Based on Limited Samples
  • space * Remote sensing cross-modal text-image retrieval based on global and local information
    space * [TGRS 2022, ESI Highly Cited, first author] [paper] [mainpage]

    space * A lightweight multi-scale crossmodal text-image retrieval method in remote sensing
    space * [TGRS 2021, first author] [paper] [mainpage]

    space * MCRN: A multi-source cross-modal retrieval network for remote sensing
    space * [JAG 2022, first author] [paper] [mainpage]

    space * Learning to Evaluate Performance of Multi-modal Semantic Localization
    space * [TGRS 2022, first author] [paper] [mainpage]

    space * SeLo v2: Toward for higher and faster semantic localization
    space * [GRSL 2023, corresponding author] [paper] [mainpage]

  • Model Robustness Design
  • space * Visual transformer with stable prior and patch-level attention for single image dehazing
    space * [Neurocomputing 2023, corresponding author] [paper]

    space * Frequency compensated diffusion model for real-scene dehazing
    space * [NeuralNetworks 2024, first author as a student] [paper] [mainpage]

    space * Diffusion Model with Detail Complement for Super-Resolution of Remote Sensing
    space * [RemoteSensing 2022, corresponding student] [paper]

    space * A Synthetic-to-Real Dehazing Method based on Domain Unification
    space * [ICME 2025, first author]

  • Dataset Construction
  • space * Exploring a fine-grained multiscale method for cross-modal remote sensing image retrieval
    space * [TGRS 2021, first student] [paper] [mainpage]

    space * VSD2M: A Large-scale Vision-language Sticker Dataset for Multi-frame Animated Sticker Generation
    space * [ICME 2025, first author] [paper] [mainpage]


    Deep Learning Under Limited Resources

  • Design of Distillation Method
  • space * ILDiff: Generate Transparent Animated Stickers by Implicit Layout Distillation
    space * [ICASSP 2025, first author] [paper] [mainpage]

    space * Efficient and Controllable Remote Sensing Fake Sample Generation Based on Diffusion Model
    space * [TGRS 2023, first author] [paper] [mainpage]

  • Lightweight Model Design and Application
  • space * WalkVLM: Aid Visually Impaired People Walking by Vision Language Model
    space * [Arxiv 2024, first author] [paper] [mainpage]

    Project