Network-Agnostic Knowledge Transfer for Medical Image Segmentation

TL;DR

This paper introduces a network-agnostic knowledge transfer method for medical image segmentation that does not require original training data or similar architectures, enabling effective transfer from multiple teachers to a student model across various datasets and modalities.

Contribution

The proposed approach allows architecture-agnostic knowledge transfer without access to original training data, facilitating ensemble learning and simplifying implementation for diverse medical imaging tasks.

Findings

Student models match or outperform teachers in segmentation performance.

Knowledge transfer is effective across different network architectures.

Method reduces dependence on original training data and generative models.

Abstract

Conventional transfer learning leverages weights of pre-trained networks, but mandates the need for similar neural architectures. Alternatively, knowledge distillation can transfer knowledge between heterogeneous networks but often requires access to the original training data or additional generative networks. Knowledge transfer between networks can be improved by being agnostic to the choice of network architecture and reducing the dependence on original training data. We propose a knowledge transfer approach from a teacher to a student network wherein we train the student on an independent transferal dataset, whose annotations are generated by the teacher. Experiments were conducted on five state-of-the-art networks for semantic segmentation and seven datasets across three imaging modalities. We studied knowledge transfer from a single teacher, combination of knowledge transfer and fine-tuning, and knowledge transfer from multiple teachers. The student model with a single teacher achieved similar performance as the teacher; and the student model with multiple teachers achieved better performance than the teachers. The salient features of our algorithm include: 1)no need for original training data or generative networks, 2) knowledge transfer between different architectures, 3) ease of implementation for downstream tasks by using the downstream task dataset as the transferal dataset, 4) knowledge transfer of an ensemble of models, trained independently, into one student model. Extensive experiments demonstrate that the proposed algorithm is effective for knowledge transfer and easily tunable.