publication . Thesis . 2017

Unsupervised Learning of Spatiotemporal Features by Video Completion

Nallabolu, Adithya Reddy;
Open Access
  • Published: 18 Oct 2017
  • Publisher: Virginia Tech
  • Country: United States
Abstract
In this work, we present an unsupervised representation learning approach for learning rich spatiotemporal features from videos without the supervision from semantic labels. We propose to learn the spatiotemporal features by training a 3D convolutional neural network (CNN) using video completion as a surrogate task. Using a large collection of unlabeled videos, we train the CNN to predict the missing pixels of a spatiotemporal hole given the remaining parts of the video through minimizing per-pixel reconstruction loss. To achieve good reconstruction results using color videos, the CNN needs to have a certain level of understanding of the scene dynamics and predi...
Subjects
arXiv: Computer Science::MultimediaComputer Science::Computer Vision and Pattern Recognition
ACM Computing Classification System: ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION
free text keywords: Representation Learning, Supervised, Unsupervised
Download from
VTechWorks
Thesis . 2017
Provider: VTechWorks
Powered by OpenAIRE Open Research Graph
Any information missing or wrong?Report an Issue