معرفی
Vinkle Srivastav is a Research Scientist (Chargé de recherche R&D) at the CAMMA group, a collaborative research team between IHU Strasbourg and the University of Strasbourg, where he focuses on advancing surgical data science through novel computer vision and machine learning approaches. His work bridges the gap between clinical practice and artificial intelligence, developing methods for surgical video analysis, 3D medical imaging, and surgical workflow understanding.
Education
- PhD in Computer Science (2018-2021) from University of Strasbourg, France. Thesis: "Unsupervised Domain Adaptation Approaches for Person Localization in the Operating Rooms."
- Master of Science in Computer Science (2014-2017) from Indian Institute of Technology, Delhi, India. Thesis: "Computerized evaluation of neurosurgery skills using image processing and computer vision techniques."
- Bachelor of Technology in Electronics and Communication (2007-2011) from Punjab Technical University, Jalandhar, India.
Research Interests
Vinkle's research spans surgical data science, with particular focus on multi-modal learning approaches for surgical computer vision. His work addresses fundamental challenges in medical AI including domain adaptation, self-supervised learning, and privacy preservation in clinical environments. He develops methods for 3D medical image analysis, multi-view human pose estimation in operating rooms, and surgical activity recognition.
His recent work emphasizes multi-modal pretraining frameworks that leverage both visual and textual information to improve surgical workflow understanding. He also investigates scientific simulation techniques, particularly for therapeutic ultrasound applications, where physics-aware deep learning models can accelerate computational processes while maintaining accuracy.
Publication Trends
Vinkle's recent publications demonstrate a strong trajectory toward multi-modal surgical AI systems that integrate vision, language, and physics-based modeling. His work increasingly focuses on few-shot and zero-shot adaptation techniques to address the data scarcity problem in surgical AI. The publications reveal a progression from basic pose estimation to holistic surgical scene understanding, incorporating team communication analysis and surgical safety protocols.
Scientific Awards
- IPCAI 2024 Best paper award (co-author)
- IPCAI 2019 Runner-up award in the bench-to-bedside category (co-author)
- Joint winner for the best paper award in the machine learning for CAI track, IPCAI 2025
Advising and Grants
Vinkle actively mentors multiple PhD students and research interns at various levels, supervising thesis work on topics including large-scale multi-modality learning, holistic surgical scene analysis, and self-supervised video representation learning. He serves as Co-PI on two ITI-HealthTech projects: one focused on multi-modality learning for 3D medical imaging (2023), and another on physics-aware deep-learning approaches for therapeutic ultrasound simulation (2024).
Laboratories and Teams
Vinkle is a key member of the CAMMA research group at IHU Strasbourg, a collaborative team focused on computer-assisted medical modeling and analytics. He co-organizes the Surgical Data Science Summer School, an interdisciplinary program that brings together clinicians and computer scientists to develop AI-driven solutions with clinical impact. His work involves close collaboration with surgical teams at University Hospitals of Strasbourg and international partners including Johns Hopkins University and Technical University of Munich.
