Cross domain segmentation with uncertainty-guided curriculum learning

#The present invention relates generally to medical image segmentation, and in particular to cross domain segmentation with uncertainty-guided curriculum learning.

#Deep neural networks have been widely applied in medical imaging analysis for performing a variety of clinical tasks. One important clinical task is medical image segmentation. Conventionally, training deep neural networks for medical image segmentation requires large amounts of annotated training data to avoid overfitting. However, annotating such training data is a difficult, expensive, and time-consuming task which requires extensive clinical expertise. Accordingly, obtaining annotated training data at a large enough scale for training deep neural networks is a substantial challenge.

#In accordance with one or more embodiments, systems and methods for training a machine learning based segmentation network are provided. A set of medical images, each depicting an anatomical object, in a first modality is received. For each respective medical image of the set of medical images, a synthetic image, depicting the anatomical object, in a second modality is generated based on the respective medical image. One or more augmented images are generated based on the synthetic image. One or more segmentations of the anatomical object are performed from the one or more augmented images using a machine learning based reference network. An uncertainty associated with segmenting the anatomical object from the respective medical image is computed based on results of the one or more segmentations. It is determined whether the respective medical image is suitable for training a machine learning based segmentation network based on the uncertainty. The machine learning based segmentation network is trained based on 1) the suitable medical images of the set of medical images and 2) annotations of the anatomical object determined using a machine learning based teacher network.

#FIG. 1 shows a framework for training a machine learning based segmentation network for medical image segmentation using uncertainty-guided curriculum learning, in accordance with one or more embodiments;

#FIG. 2 shows a method for training a segmentation network for medical image segmentation, in accordance with one or more embodiments;

#FIG. 3 shows a workflow for unsupervised training of a segmentation network for medical image segmentation, in accordance with one or more embodiments;

#The present invention generally relates to methods and systems for cross domain segmentation with uncertainty-guided curriculum learning. Embodiments of the present invention are described herein to give a visual understanding of such methods and systems. A digital image is often composed of digital representations of one or more objects (or shapes). The digital representation of an object is often described herein in terms of identifying and manipulating the objects. Such manipulations are virtual manipulations accomplished in the memory or other circuitry/hardware of a computer system. Accordingly, is to be understood that embodiments of the present invention may be performed within a computer system using data stored within the computer system. Embodiments of the present invention will be described herein with reference to the drawings, where like reference numerals represent the same or similar elements.

#Embodiments described herein provide for a cross domain framework for training a machine learning based segmentation network for medical image segmentation using uncertainty-guided curriculum learning. FIG. 1 shows a framework 100 for training a machine learning based segmentation network for medical image segmentation using uncertainty-guided curriculum learning, in accordance with one or more embodiments. Framework 100 comprises a two-stage approach for progressively training the segmentation network using a large-scale training dataset of unannotated (i.e., unlabeled) images.

#During the first stage, an uncertainty associated with segmenting an anatomical object from each image of an unlabeled dataset 102 is determined. The anatomical objects are segmented from each image of unlabeled dataset 102 by reference network 104. Uncertainty 106 associated with the segmentation of the anatomical objects from each respective image of unlabeled data 102 is quantified and it is determined whether the respective image is considered an easy case (i.e., the respective image is suitable for training the segmentation network) or a hard case (i.e., the respective image is not suitable training the segmentation network) based on the uncertainty.