Pranav Mantini

Dept. of Computer Science, UH

pranav_profile.jpg

Senior Researcher & Lecturer

Office: PGH 524

University of Houston

Houston, TX

I’m a Senior Researcher and Lecturer at the University of Houston, where I’ve been teaching and researching since 2017. My research spans Computer Vision, Image Processing, and Machine Learning.

I design intelligent tools for public safety video systems, bridging theoretical research with robust, real-world deployment. My research addresses foundational challenges in person re-identification, anomaly detection, and neural network architectures. My work has been recognized with a U.S. Patent, best-paper awards, and significant federal grant funding.

Currently, I’m excited about exploring geometries and frameworks for parameter-efficient fine-tuning of Contrastive Vision Language Models.

news

Mar 27, 2026 🚀 New Paper Alert! Our research on BiCLIP: Bilinear CLIP for Parameter-Efficient Fine-Tuning has been accepted at DGEBF 2026, a CVPR workshop. This work introduces a bilinear unit to align vision-language manifolds with minimal parameters.
Mar 26, 2026 Check out the project page of Bilinear CLIP for more results and visualization. URL
Feb 04, 2026 Downlaod Imagenet-16 a subset of Imagenet-1k for few-shot classification. URL

selected publications

  1. BiCLIP: Domain Canonicalization via Structured Geometric Transformation
    Pranav Mantini and Shishir K Shah
    arXiv preprint arXiv:2603.08942, 2026
  2. Tackling domain shifts in person re-identification: A survey and analysis
    Vuong D Nguyen, Samiha Mirza, Abdollah Zakeri, and 6 more authors
    In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2024
  3. Cqnn: Convolutional quadratic neural networks
    Pranav Mantini and Shishr K Shah
    In 2020 25th International Conference on Pattern Recognition (ICPR), 2021