ReShapeIT: Reliable Shape Interaction with Implicit Template for Medical Anatomy Reconstruction

By Minghui Zhang, and Yun Gu

Institute of Medical Robotics, Shanghai Jiao Tong University

We present the Reliable Shape Interaction with Implicit Template (ReShapeIT) network, which represents anatomical structures using continuous implicit fields rather than discrete voxel grids. The approach combines a category-specific implicit template field with a deformation field to encode anatomical shapes from training shapes. In addition, a Template Interaction Module (TIM) is designed to refine test cases by aligning learned template shapes with instance-specific latent codes.

Detailed environment configurations are listed in the file environment.yaml.

Pretrained model weights are available in the folder checkpoint_models.

To train the model, run:

python train.py

To fine-tune the model with the specified configuration files, run:

python finetune.py

Detailed configurations are provided in the folder configs.