7-segmentation of 200 scenes from BDD100k

7-segmentation of 200 scenes from BDD100k

K-Segmentation of a video stream is used to partition it into k-piecewise-linear segments. Each of these segments is expected to have a meaningful interpretation. This can be used in summarizing videos, identifying anomalies, and selecting critical subsets for machine learning.K-Segmentation is demo...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Sabarish Vadarevu, Karamcheti, Vijay
Format: Video
Sprache:eng
Schlagworte:
FOS: Computer and information sciences
FOS: Psychology
Knowledge Representation and Machine Learning
Pattern Recognition and Data Mining
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:K-Segmentation of a video stream is used to partition it into k-piecewise-linear segments. Each of these segments is expected to have a meaningful interpretation. This can be used in summarizing videos, identifying anomalies, and selecting critical subsets for machine learning.K-Segmentation is demonstrated here for 200 scenes sampled from the Berkeley DeepDrive dataset, using four different algorithms. The segmentations are marked by using different border colors (cycling R/G/B) for each segment.
DOI:10.6084/m9.figshare.12859493