Abstract
Image segmentation algorithms for medical embedded vision devices usually require a light and low latency model. In this study, a novel lightweight DepthWise U-shape network (DWU-net) is proposed to address this issue, which implements the task of lung parenchyma image segmentation. In the contracting path and expanding path of segmentation network, we introduce a separable convolution unit to replace standard convolution for image feature extraction, which can learn the unique features of each layer of the image from multiple perspectives, and has more advantages in feature expression. Our algorithm has better flexibility, comparing to the original model, the model parameters’ number has been greatly reduced and the time efficiency is fully improved. The proposed architecture achieves good performance in FPGA implementation.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: International Journal of Pattern Recognition and Artificial Intelligence
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
