LmPT: Conditional Point Transformer for Anatomical Landmark Detection on 3D Point Clouds

TL;DR

The paper tackles automated anatomical landmark detection on 3D point clouds, addressing inter-observer variability and cross-species generalization. It introduces LmPT, a conditional Point Transformer with FiLM modulation that conditions the model on input type to enable learning across homologous bones, here demonstrated on human and dog femurs with a new dog dataset. The approach achieves state-of-the-art landmark localization (MAE and PCK) and shows that cross-species training can improve performance for humans while maintaining competitive results for dogs, enabling translational research. This work highlights the practicality of point-cloud methods for scalable, accurate anatomical landmarking and provides a dataset to foster broader cross-species studies.

Abstract

Accurate identification of anatomical landmarks is crucial for various medical applications. Traditional manual landmarking is time-consuming and prone to inter-observer variability, while rule-based methods are often tailored to specific geometries or limited sets of landmarks. In recent years, anatomical surfaces have been effectively represented as point clouds, which are lightweight structures composed of spatial coordinates. Following this strategy and to overcome the limitations of existing landmarking techniques, we propose Landmark Point Transformer (LmPT), a method for automatic anatomical landmark detection on point clouds that can leverage homologous bones from different species for translational research. The LmPT model incorporates a conditioning mechanism that enables adaptability to different input types to conduct cross-species learning. We focus the evaluation of our approach on femoral landmarking using both human and newly annotated dog femurs, demonstrating its generalization and effectiveness across species. The code and dog femur dataset will be publicly available at: https://github.com/Pierreoo/LandmarkPointTransformer.

Figures (7)

• Figure 1: Ground truth landmarks for human and dog femurs.
• Figure 2: Overview of femur processing with the PTv2 encoder.
• Figure 3: Overview of LmPT, outlining a PT encoder-decoder structure with a FiLM modulation to condition the model by input.
• Figure 4: Human (left) and dog (right) femur dataset, PCK (%).
• Figure 5: Method comparison of human landmark predictions.