eduzhai > Applied Sciences > Engineering >

Sensorless Freehand 3D Ultrasound Reconstruction via Deep Contextual Learning

  • Save

... pages left unread,continue reading

Document pages: 10 pages

Abstract: Transrectal ultrasound (US) is the most commonly used imaging modality toguide prostate biopsy and its 3D volume provides even richer contextinformation. Current methods for 3D volume reconstruction from freehand USscans require external tracking devices to provide spatial position for everyframe. In this paper, we propose a deep contextual learning network (DCL-Net),which can efficiently exploit the image feature relationship between US framesand reconstruct 3D US volumes without any tracking device. The proposed DCL-Netutilizes 3D convolutions over a US video segment for feature extraction. Anembedded self-attention module makes the network focus on the speckle-richareas for better spatial movement prediction. We also propose a novel case-wisecorrelation loss to stabilize the training process for improved accuracy.Highly promising results have been obtained by using the developed method. Theexperiments with ablation studies demonstrate superior performance of theproposed method by comparing against other state-of-the-art methods. Sourcecode of this work is publicly available atthis https URL.

Please select stars to rate!


0 comments Sign in to leave a comment.

    Data loading, please wait...