Validation

We validate our CathSim simulator and robots through testing on both phantom models and animal subjects.

Phantom SetupPermalink

xray_room_slave

To simulate human vascular anatomy, we use a half-body vascular phantom model, which is connected to a closed water circuit to mimic blood flow. We utilize a Bi-planar X-ray system equipped with 60 kW Epsilon X-ray Generators and 16-inch Image Intensifier Tubes for high-definition imaging.

xray_room_master

The surgeon operates the master device from a control room to drive the follower robot located in the X-ray room.

Animal SetupPermalink

animal_setup

We use pigs as animal subjects for our validation, as their anatomy shares similarities with that of humans.

Catheterization TargetPermalink

setup_exp

The catheterization targets are the brachiocephalic artery (BCA) and the left common carotid artery (LCCA).

SplineFormer NetworkPermalink

splineformer

We propose SplineFormer, a new transformer-based architecture, designed specifically to predict the continuous, smooth shape of the guidewire in an explainable way.

Expert Navigation NetworkPermalink

expert_network

Our Expert Navigation Network is a multimodal system trained using CathSim simulator and subsequently transferred to the real robot.

Segmentation ModelPermalink

segmentation_model

We introduce a shape-sensitive loss function for catheter and guidewire segmentation and utilize it in a vision transformer network to establish a new state-of-the-art result on a large-scale X-ray images dataset.

ResultsPermalink

expert_result
expert_table

The results demonstrate the successful integration of simulation, machine learning, and vision technologies to achieve autonomous catheterization.

Updated: