Lennart Bastian

I am a Research Fellow at Imperial College London, supported through the UK Royal Society Newton International Fellowship. I work with the Circle Group of Prof. Tolga Birdal, where I pursue topics at the intersection of geometry, topology, and machine learning as a Post-doc.

I completed my PhD with highest distinction (summa cum laude) at the Chair for Computer Aided Medical Procedures (CAMP) in Munich, supervised by Prof. Nassir Navab, where I also coordinated research of the Surgical Data Science team. Prior to this, I conducted my M.Sc. in Applied Mathematics at the Technical University of Munich with an emphasis on optimization and statistics, and Bachelor's in Mathematics and Computer Science from New York University's Courant Institute.

I also had previous stints in software and data science at H2oMetrics, a cloud water management startup, and as a research assistant at the MSKCC Levine Lab, developing computational tools for the genetic analysis of patients with myeloid leukemias.

Email  /  LinkedIn  /  Google Scholar  /  GitHub

Selected Publications

My research interests span the application of ideas from geometry and topology in machine learning and computer vision, including application areas in medicine and biology, and shape analysis.

Copresheaf Topological Neural Networks: A Generalized Deep Learning Framework

We introduce copresheaf topological neural networks (CTNNs), a powerful and unifying framework that encapsulates a wide spectrum of deep learning architectures, designed to operate on structured data. The framework aims to address challenges in representation learning by using concepts from algebraic topology.

Forecasting Continuous Non-Conservative Dynamical Systems in SO(3)

We study rotational forecasting on the manifold SO(3). This leads to the proposed Savitzky-Golay Neural Controlled Differential Equations which learn continuous data-driven priors on the manifold and excel at extrapolating rotational states.

Beyond Role-Based Surgical Domain Modeling

We achieve generalizable re-identification and tracking for surgical operating rooms (OR). This enables our proposed staff-centric surgical domain models, the first concept for personalized intelligent systems in the OR.

Beyond Complete Shapes: A Quantitative Evaluation of 3D Shape Matching Algorithms

We present BeCoS, the first comprehensive benchmark and evaluation framework for the challenging but widely applicable problem setting of partial shape correspondence.

Hybrid Functional Maps for Crease-Aware Non-Isometric Shape Matching

We propose a hybrid approach for estimating functional maps that leverage the strengths of basis functions originating from different operators overcoming limitations of previous approaches for non-isometric deformations.

S3M: Scalable Statistical Shape Modeling through Unsupervised Correspondences

We propose an unsupervised method for statistical shape model (SSM) construction that utilizes functional correspondences to learn shape structures across population anatomies, demonstrating robustness to noisy and potentially scaling to larger patient populations without manual annotations.

SegmentOR: Obtaining Efficient Operating Room Semantics through Temporal Propagation

SegmentOR is a weakly-supervised 3D semantic segmentation method for operating room environments that leverages temporal consistency in 4D point cloud sequences to significantly reduce the annotation burden. We demonstrate the utility of the resulting segmentation maps for improving downstream surgical workflow analysis.

DisguisOR: Holistic Face Anonymization for the Operating Room

DisguisOR improves the robustness of operating room video anonymization by leveraging multi-view data to accurately localize and anonymize individuals’ faces in 3D, resulting in geometrically consistent privacy protection across all camera views.

Preprints & Workshops

COMPOSE: Hypergraph Cover Optimization for Multi-view 3D Human Pose Estimation

We reformulate multi-view 3D pose estimation as a hypergraph partitioning problem, introducing efficient geometric pruning to achieve up to 23% improvement in average precision over previous optimization-based methods.

Mutual Information Free Topological Generalization Bounds via Stability

We develop a new theoretical framework using algorithmic stability and topological data analysis to establish generalization bounds without complex mutual information calculations.

Mitigating Biases in Surgical Operating Rooms with Geometry

We show that 3D point cloud sequences capture robust biometric information for surgical personnel recognition, avoiding the appearance-based shortcuts that cause RGB models to fail in real clinical settings.

TrackOR: Towards Personalized Intelligent Operating Rooms Through Robust Tracking

We present TrackOR, a robust tracking framework for personalized intelligent operating rooms that enables consistent personnel identification across surgical procedures.

Design adapted from here