Hao Ju (鞠豪 in Chinese)

I am a first-year second-year third-year Ph.D. student at University of Macau, supervised by Dr. Shaofei Huang, Dr. Hu Zhang, and Prof. Zhedong Zheng. Previously, I received my Bachelor’s and Master’s degrees from Dalian University of Technology in 2021 and 2024, supervised by Dr. Pengyu Zhang, Prof. Xu Jia and Prof. Dong Wang.

I am interested in Computer Vision, Embodied AI and related topics.

📧 Email / 🎓 Scholar

🌟 Highlight: I am currently seeking a long‑term research internship opportunity. Please feel free to contact me if you have an opportunity!

 Profile Photo

News

  • 2026‑01: Passed PhD Qualifying Examination @ UM.
  • 2026‑01: (First author) One paper on event-based lip reading is accepted by ACM TOMM.
  • 2025‑06: (First author) One paper on cross‑view geo‑localization is accepted by ICCV 2025.
  • 2024‑09: (Equal contribution) One paper on event‑assisted blurry image unfolding is accepted by IEEE TNNLS.
  • 2024‑07: (Equal contribution) One paper on event‑assisted blurry image unfolding is accepted by IEEE TIP.

Research

MuSTMM: Learning Sound-Triggered Mobile Manipulation across Multiple Scenes
Authors: Hao Ju*, Feng Yan*, Zhihao Zhong, Heng Yang, Shaofei Huang, Zhedong Zheng
Under Review (* denotes equal contribution)
Click the image on the left for a video demo. (Headphones recommended)
Paper  /
We introduce MuSTMM, a multi-scene STMM benchmark, and augment sound navigation with explicit spatial memory for unseen scenes.
From Instruction to Event: Sound-Triggered Mobile Manipulation
Authors: Hao Ju, Shaofei Huang, Hongyu Li, Zihan Ding, Si Liu, Meng Wang, Zhedong Zheng
Under Review
Click the image on the left for a video demo. (Headphones recommended)
Paper  /
Instruction-following alone is not enough. We propose a paradigm shift to event-triggered robots that ground sound events and respond autonomously.
tfnet placeholder Event-based Lip Reading with Triplane Fusion Network
Authors: Hao Ju, Zhedong Zheng, Xu Zheng, Wenyue Chen, Lin Wang, Dong Wang, Huchuan Lu, Xu Jia
ACM TOMM
Paper  /
We leverage the high temporal resolution of event data and process event stream in a triplane manner.
Video2BEV placeholder Video2BEV: Transforming Drone Videos to BEVs for Video‑based Geo‑localization
Authors: Hao Ju*, Shaofei Huang*, Si Liu, Zhedong Zheng
ICCV 2025 (* denotes equal contribution)
Paper  /  Code
We resort to the multi‑view nature of drone‑view videos and transform videos to Bird’s‑Eye View for video‑based geo‑localization.
EARN placeholder Event‑Assisted Recurrent Network for Arbitrary‑Temporal‑Scale Blurry Image Unfolding
Authors: Pengyu Zhang*, Hao Ju*, Weihua He, Yaoyuan Wang, Ziyang Zhang, Shengming Li, Dong Wang, Huchuan Lu, Xu Jia
IEEE TNNLS 2024 (* denotes equal contribution)
PDF /
We propose an event‑assisted blurry image unfolding framework that can work across arbitrary temporal scales with the help of event streams.
EBRL placeholder Event‑Assisted Blurriness Representation Learning for Blurry Image Unfolding
Authors: Pengyu Zhang*, Hao Ju*, Lei Yu, Weihua He, Yaoyuan Wang, Ziyang Zhang, Qi Xu, Shengming Li, Dong Wang, Huchuan Lu, Xu Jia
IEEE TIP 2024 (* denotes equal contribution)
PDF /
We propose to implicitly model blur in an image by computing blurriness representation from event streams and integrate blurriness into the unfolding network.

Education

  • 2024‑09 – Now: Ph.D. student, Computer Science, University of Macau
  • 2021‑06 – 2024‑06: M.S., Information & Communication Engineering, Dalian University of Technology
  • 2017‑09 – 2021‑06: B.S., Information & Communication Engineering, Dalian University of Technology

Interests

Pick & Roll

Catch & Shoot

Two Ball Dribbling

Logo Shot

Visitor Statistics

Template from Jon Barron's website.