A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Filters
MixModule: Mixed CNN Kernel Module for Medical Image Segmentation
[article]
2020
arXiv
pre-print
Preserved Fulltext
Other Versions
Convolutional neural networks (CNNs) have been successfully applied to medical image classification, segmentation, and related tasks. ...
In this paper, we propose a module that combines the benefits of multiple kernel sizes and we apply the proposed module to U-Net and its variants. ...
PROPOSED METHOD
MixModule MixModule contains multiple sizes of convolution kernels to capture different ranges of semantic information which is crucial for medical images that emphasize the details of ...
arXiv:1910.08728v2
fatcat:q2vi2ct62neidbwfhtlaq7k3he