Jitesh Joshi

Researcher @ University College London (UCL).

prof_pic.jpg

Desk: 3.05

169 Euston Road

London, UK, NW1 2AE

UCL Profile Page

I’m a PhD researcher in Computer Science at University College London (UCL), specializing in robust computer vision and deep learning systems that perform reliably across diverse real-world conditions. My research advances contactless camera-based physiological sensing through novel multidimensional attention mechanisms grounded in non-negative matrix factorization, enabling joint attention computation across spatial, temporal, and channel dimensions.

This work, published at NeurIPS 2024, achieved significant improvements in cross-dataset generalization while maintaining computational efficiency—addressing a critical challenge in deploying AI systems beyond controlled laboratory settings. My research spans computer vision, foundational models, physiological computing, and generative AI, with particular focus on architectures that balance robustness with efficiency.

Prior to my PhD, I spent over a decade in healthcare technology, leading AI, R&D, and systems engineering teams to translate research prototypes into commercial products, including FDA/CE-certified medical devices. This experience bridging research and product development informs my approach to tackling technically fascinating challenges while ensuring meaningful societal impact.

Some of my recent research publications are highlighted below.

news

Jul 2, 2025 I am delighted to announce that I have successfully passed my final PhD viva! This milestone marks the culmination of an incredible research journey that began in October 2020. I extend my deepest gratitude to my PhD supervisors, Prof. Youngjun Cho and Prof. Nadia Berthouze, whose guidance and mentorship have been invaluable throughout this journey. I would also like to express my sincere appreciation to my examiners, Prof. Anthony Steed and Prof. Yannick Benezeth, for their thoughtful engagement and constructive feedback. Finally, I am deeply grateful to all the colleagues, friends, and family who have supported me along the way—this achievement would not have been possible without your encouragement and support.
Jun 2, 2025 How matrix factorization can serve as attention mechanism? Check out our blog on our NeurIPS, 2024 work, titled “FactorizePhys: Matrix Factorization for Multidimensional Attention in Remote Physiological Sensing” that establishes strong baseline for cross-dataset generalization in remote physiological sensing.
May 18, 2025 Check out the web-browser based real-time demo of our recent work on multitask estimation of remote photoplethysmography (rPPG) and remote respiratory (rRSP) signals from facial video. Preprint of this work can be accessed on arXiv.
May 12, 2025 Today marks a significant milestone as I submitted my PhD thesis, “Enhancing Out-of-distribution Generalization for Robust Camera-based Remote Physiological Sensing,” concluding four years of research at University College London with defense scheduled for July 2025.
Apr 15, 2024 iBVP Dataset: RGB-Thermal rPPG dataset with signal quality labels, MDPI Sensors, 2024

selected publications

  1. MMRPhys.png
    Efficient and Robust Multidimensional Attention in Remote Physiological Sensing through Target Signal Constrained Factorization
    Jitesh Joshi, and Youngjun Cho
    arXiv: 2505.07013 [cs.CV], 2025
  2. FSAM.png
    FactorizePhys: Matrix Factorization for Multidimensional Attention in Remote Physiological Sensing
    Jitesh Joshi, Sos Agaian, and Youngjun Cho
    In The Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024
  3. iBVP_dataset.png
    iBVP Dataset: RGB-Thermal rPPG Dataset with High Resolution Signal Quality Labels
    Jitesh Joshi, and Youngjun Cho
    Electronics, 2024
  4. samcl.png
    Self-adversarial Multi-scale Contrastive Learning for Semantic Segmentation of Thermal Facial Images
    Jitesh Joshi, Nadia Bianchi-Berthouze, and Youngjun Cho
    In 33rd British Machine Vision Conference 2022, BMVC 2022, London, UK, November 21-24, 2022, 2022