Abstract
The automated segmentation of biomedical images is being used increasingly for the clinical practices. Deep learning techniques and networks demonstrate promising results in applications of computer vision. However, the behaviour of different networks are effectively influenced by the choice of architecture and training parameter settings. Ensembles of neural networks are known to be much more efficient and accurate than individual networks. We explore stacking ensemble and snapshot ensembles of different neural network models and architectures for robust performance through the aggregation of predictions for effective biomedical image segmentation tasks. The network models used are inspired from the Fully Convolutional Network(FCN) architectures. The proposed approach is a generic and unbiased deep learning approach that reduces the risk of over fitting the network configuration to a particular dataset. We evaluated our approach on the Boston University-Biomedical Image Library(BU-BIL) Dataset 1. The predicted image segmentation masks were compared to the gold standard ground truth annotations of the dataset to validate the effectiveness of our proposed ensemble architectures.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
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.