Show simple item record

dc.contributor.authorLahoud, Jean
dc.contributor.authorGhanem, Bernard
dc.date.accessioned2020-02-26T09:07:38Z
dc.date.available2020-02-26T09:07:38Z
dc.date.issued2020-02-06
dc.identifier.urihttp://hdl.handle.net/10754/661709
dc.description.abstractAlthough well-known large-scale datasets, such as ImageNet, have driven image understanding forward, most of these datasets require extensive manual annotation and are thus not easily scalable. This limits the advancement of image understanding techniques. The impact of these large-scale datasets can be observed in almost every vision task and technique in the form of pre-training for initialization. In this work, we propose an easily scalable and self-supervised technique that can be used to pre-train any semantic RGB segmentation method. In particular, our pre-training approach makes use of automatically generated labels that can be obtained using depth sensors. These labels, denoted by HN-labels, represent different height and normal patches, which allow mining of local semantic information that is useful in the task of semantic RGB segmentation. We show how our proposed self-supervised pre-training with HN-labels can be used to replace ImageNet pre-training, while using 25x less images and without requiring any manual labeling. We pre-train a semantic segmentation network with our HN-labels, which resembles our final task more than pre-training on a less related task, e.g. classification with ImageNet. We evaluate on two datasets (NYUv2 and CamVid), and we show how the similarity in tasks is advantageous not only in speeding up the pre-training process, but also in achieving better final semantic segmentation accuracy than ImageNet pre-training
dc.publisherarXiv
dc.relation.urlhttps://arxiv.org/pdf/2002.02200
dc.rightsArchived with thanks to arXiv
dc.titleRGB-based Semantic Segmentation Using Self-Supervised Depth Pre-Training
dc.typePreprint
dc.contributor.departmentComputer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division
dc.contributor.departmentElectrical Engineering Program
dc.contributor.departmentVCC Analytics Research Group
dc.eprint.versionPre-print
dc.identifier.arxivid2002.02200
kaust.personLahoud, Jean
kaust.personGhanem, Bernard
refterms.dateFOA2020-02-26T09:08:28Z


Files in this item

Thumbnail
Name:
Preprintfile1.pdf
Size:
7.957Mb
Format:
PDF
Description:
Pre-print

This item appears in the following Collection(s)

Show simple item record