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/

Zendy Plus

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

Zendy Tools

Zendy Open

Scene flow is a challenging task aimed at jointly estimating the 3D structureand motion of the sensed environment. Although deep learning solutions achieveoutstanding performance in terms of accuracy, these approaches divide the wholeproblem into standalone tasks (stereo and optical flow) addressing them withindependent networks. Such a strategy dramatically increases the complexity ofthe training procedure and requires power-hungry GPUs to infer scene flowbarely at 1 FPS. Conversely, we propose DWARF, a novel and lightweightarchitecture able to infer full scene flow jointly reasoning about depth andoptical flow easily and elegantly trainable end-to-end from scratch. Moreover,since ground truth images for full scene flow are scarce, we propose toleverage on the knowledge learned by networks specialized in stereo or flow,for which much more data are available, to distill proxy annotations.Exhaustive experiments show that i) DWARF runs at about 10 FPS on a singlehigh-end GPU and about 1 FPS on NVIDIA Jetson TX2 embedded at KITTI resolution,with moderate drop in accuracy compared to 10x deeper models, ii) learning frommany distilled samples is more effective than from the few, annotated onesavailable. Code available at:https://github.com/FilippoAleotti/Dwarf-Tensorflow

Learning End-To-End Scene Flow by Distilling Single Tasks Knowledge