eduzhai > Applied Sciences > Engineering >

Suggestive Annotation of Brain Tumour Images with Gradient-guided Sampling

  • Save

... pages left unread,continue reading

Document pages: 10 pages

Abstract: Machine learning has been widely adopted for medical image analysis in recentyears given its promising performance in image segmentation and classificationtasks. As a data-driven science, the success of machine learning, in particularsupervised learning, largely depends on the availability of manually annotateddatasets. For medical imaging applications, such annotated datasets are noteasy to acquire. It takes a substantial amount of time and resource to curatean annotated medical image set. In this paper, we propose an efficientannotation framework for brain tumour images that is able to suggestinformative sample images for human experts to annotate. Our experiments showthat training a segmentation model with only 19 suggestively annotated patientscans from BraTS 2019 dataset can achieve a comparable performance to traininga model on the full dataset for whole tumour segmentation task. It demonstratesa promising way to save manual annotation cost and improve data efficiency inmedical imaging applications.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×