English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

We investigate the effectiveness of a simple solution to the common problemof deep learning in medical image analysis with limited quantities of labeledtraining data. The underlying idea is to assign artificial labels to abundantlyavailable unlabeled medical images and, through a process known as surrogatesupervision, pre-train a deep neural network model for the target medical imageanalysis task lacking sufficient labeled training data. In particular, weemploy 3 surrogate supervision schemes, namely rotation, reconstruction, andcolorization, in 4 different medical imaging applications representingclassification and segmentation for both 2D and 3D medical images. 3 keyfindings emerge from our research: 1) pre-training with surrogate supervisionis effective for small training sets; 2) deep models trained from initialweights pre-trained through surrogate supervision outperform the same modelswhen trained from scratch, suggesting that pre-training with surrogatesupervision should be considered prior to training any deep 3D models; 3)pre-training models in the medical domain with surrogate supervision is moreeffective than transfer learning from an unrelated domain (e.g., naturalimages), indicating the practical value of abundant unlabeled medical imagedata.

Surrogate Supervision for Medical Image Analysis: Effective Deep Learning From Limited Quantities of Labeled Data