Interactive Medical Image Segmentation Using Deep Learning With Image-Specific Fine Tuning

TLDR

Convolutional neural networks excel at automatic medical image segmentation but fall short in clinical accuracy, robustness, and adaptability to unseen object classes. The authors propose an interactive segmentation framework that integrates CNNs with bounding box and scribble inputs to overcome these limitations. They fine‑tune a CNN on each test image—either unsupervised or with scribbles—using a weighted loss that accounts for network and interaction uncertainty, and apply this approach to fetal MR organ segmentation and brain tumor segmentation with sparse training labels. Experiments show the method is more robust to unseen objects, significantly improves segmentation accuracy through image‑specific fine tuning, and requires fewer user interactions and less time than conventional interactive segmentation.

Abstract

Convolutional neural networks (CNNs) have achieved state-of-the-art performance for automatic medical image segmentation. However, they have not demonstrated sufficiently accurate and robust results for clinical use. In addition, they are limited by the lack of image-specific adaptation and the lack of generalizability to previously unseen object classes (a.k.a. zero-shot learning). To address these problems, we propose a novel deep learning-based interactive segmentation framework by incorporating CNNs into a bounding box and scribble-based segmentation pipeline. We propose image-specific fine tuning to make a CNN model adaptive to a specific test image, which can be either unsupervised (without additional user interactions) or supervised (with additional scribbles). We also propose a weighted loss function considering network and interaction-based uncertainty for the fine tuning. We applied this framework to two applications: 2-D segmentation of multiple organs from fetal magnetic resonance (MR) slices, where only two types of these organs were annotated for training and 3-D segmentation of brain tumor core (excluding edema) and whole brain tumor (including edema) from different MR sequences, where only the tumor core in one MR sequence was annotated for training. Experimental results show that: 1) our model is more robust to segment previously unseen objects than state-of-the-art CNNs; 2) image-specific fine tuning with the proposed weighted loss function significantly improves segmentation accuracy; and 3) our method leads to accurate results with fewer user interactions and less user time than traditional interactive segmentation methods.