Allen Tu

Portrait

atu1@umd.edu



About Me

I am a third-year graduate student in Computer Science at the University of Maryland, College Park advised by Professor Tom Goldstein. I transferred to the PhD program in Spring 2025 after completing my BS/MS in Computer Science in 2024.

I am broadly interested in the intersection of deep learning, computer vision, and graphics. My recent work focuses on 3D scene reconstruction, and I am currently funded by the IARPA Walk-through Rendering from Images of Varying Altitude (WRIVA) program. Additionally, I have research and teaching experience in multimodal biometric recognition, large language models, vision encoders, generative adversarial networks, image quality metrics, and visualization recommendation.

News

Research Highlights



Achieve 10.38× faster rendering, 7.71× smaller model size, and 2.71× shorter training time for dynamic 3D Gaussian Splatting representations through temporal sensitivity pruning and flow-based Gaussian grouping while maintaining visual quality.





Accelerate 3D Gaussian Splatting rendering speed by over 6× and reduce model size by over 90% through accurately localizing primitives during rasterization and pruning the scene during training, providing a significantly higher speedup than existing techniques while maintaining competitive image quality.





Prune 90% of primitives from any pretrained 3D Gaussian Splatting model using a mathematically principled sensitivity score, more than tripling rendering speed while retaining more salient foreground information and higher visual fidelity than previous techniques at a substantially higher compression ratio.



Zoomable Image

Predict and Aggregate [1, 2] improves the CMC and ROC performance of the SOTA face recognition system by an empirical upper bound of 11.54% at the evaluation thresholds.


Biometric Recognition and Identification at Altitude and Range (IARPA BRIAR)
Systems & Technology Research (STR), 2022-2025.

I researched face recognition, body recognition, and multimodal fusion for our SOTA system. Ask me about:
  1. Recognizability prediction for face, body, and gait recognition. (2025)
  2. Knowledge distillation for face probe-to-gallery feature similarity prediction. (2024)
  3. Feature clustering and aggregation for face recognition from videos. (2024)
  4. Training operating condition-invariant face encoders. (2023)
  5. Model ensembling and mixed-voting classifiers for open search. (2023)
  6. Data augmentation via garment transfer for training clothing-invariant body encoders. (2022)

* denotes equal contribution.

Research Experience

Systems & Technology Research (STR)
Video and Image Understanding Group
  • Peer Research Mentor: January 2021 — December 2022
    Advisor: Dr. Raymond Tu