NCAdapt: Dynamic adaptation with domain-specific Neural Cellular Automata for continual hippocampus segmentation

TL;DR

NCAdapt introduces a domain-specific Neural Cellular Automata approach for continual hippocampus segmentation, effectively adapting to new domains with minimal parameters and achieving state-of-the-art results in medical imaging CL tasks.

Contribution

The paper presents NCAdapt, a novel NCA-based continual learning method with domain-specific adaptable layers, improving hippocampus segmentation across multiple domains.

Findings

Achieves state-of-the-art performance on hippocampus segmentation tasks.

Uses only 384 parameters for adaptation per new domain.

Outperforms Lifelong nnU-Net and U-Net with SOTA CL methods.

Abstract

Continual learning (CL) in medical imaging presents a unique challenge, requiring models to adapt to new domains while retaining previously acquired knowledge. We introduce NCAdapt, a Neural Cellular Automata (NCA) based method designed to address this challenge. NCAdapt features a domain-specific multi-head structure, integrating adaptable convolutional layers into the NCA backbone for each new domain encountered. After initial training, the NCA backbone is frozen, and only the newly added adaptable convolutional layers, consisting of 384 parameters, are trained along with domain-specific NCA convolutions. We evaluate NCAdapt on hippocampus segmentation tasks, benchmarking its performance against Lifelong nnU-Net and U-Net models with state-of-the-art (SOTA) CL methods. Our lightweight approach achieves SOTA performance, underscoring its effectiveness in addressing CL challenges in medical imaging. Upon acceptance, we will make our code base publicly accessible to support reproducibility and foster further advancements in medical CL.