Ignas Budvytis: Label propagation - semantic segmentation of 360° videos - autonomous driving scenes @EngineeringCambridge
Ignas Budvytis: Label propagation - semantic segmentation of 360° videos - autonomous driving scenes @EngineeringCambridge
Uploaded November 2017 | Updated September 2026, 1 day ago
Shortlisted in the ZEISS Photography Competition 2017.

This video illustrates labelled object classes (e.g. road sign, pavement, car etc) of a 360-degree video of Regent Street, Cambridge. The top image shows object class labels (bottom image) overlayed over a 360-image frame. Such labelled data is obtained by temporal propagation of several hand labelled video frames and is used to train neural networks for automatic object class segmentation in autonomous driving scenarios.
Ignas Budvytis: Label propagation - semantic segmentation of 360° videos - autonomous driving scenesAutomated weaving process for smart textilesThe winners – 2018 ZEISS Photography CompetitionNew augmented reality head mounted display offers unrivalled viewing experienceReconstructing a historic cricket bowling machine (part 1)The portable 3D bioprinter joining the fight against cancerUndergraduate internships at Concrete CanvasBryn Noel Ubald: An instrumented turbine blade – Another unique perspective?Detecting and tracking individual people in a crowdJonathan Harris: Shrinking SteelDr Dante McGrath: Breaking the wall of the melting ArcticVictor Serdio: Liftoff on rewind
Department of Engineering, University of Cambridge |

Ignas Budvytis: Label propagation - semantic segmentation of 360° videos - autonomous driving scenes

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER