Dataset augmentation with synthetic images improves semantic segmentation

Show simple item record Rajpura, Param Goyal, Manik Hegde, Ravi S. Bojinov, Hristo 2017-09-18T10:45:48Z 2017-09-18T10:45:48Z 2017-09
dc.identifier.citation Rajpura, Param S.; Goyal, Manik; Hegde, Ravi S. and Bojinov, Hristo, "Dataset augmentation with synthetic images improves semantic segmentation", arXiv, Cornell University Library, DOI: arXiv:1709.00849, Sep. 2017. en_US
dc.description.abstract Although Deep Convolutional Neural Networks trained with strong pixel-level annotations have significantly pushed the performance in semantic segmentation, annotation efforts required for the creation of training data remains a roadblock for further improvements. We show that augmentation of the weakly annotated training dataset with synthetic images minimizes both the annotation efforts and also the cost of capturing images with sufficient variety. Evaluation on the PASCAL 2012 validation dataset shows an increase in mean IOU from 52.80% to 55.47% by adding just 100 synthetic images per object class. Our approach is thus a promising solution to the problems of annotation and dataset collection.
dc.description.statementofresponsibility by Param S. Rajpura, Manik Goyal, Ravi S. Hegde and Hristo Bojinov
dc.language.iso en en_US
dc.publisher Cornell University Library en_US
dc.title Dataset augmentation with synthetic images improves semantic segmentation en_US
dc.type Preprint en_US

Files in this item

Files Size Format View

There are no files associated with this item.

This item appears in the following Collection(s)

Show simple item record

Search Digital Repository


My Account