Watching Unlabeled Video Helps Learn New Human Actions from Very Few Labeled Snapshots

By Chao-Yeh Chen, Kristen Grauman. 2013

42 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Representation & Self-Supervised Learning, Human-Computer Interaction

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Representation & Self-Supervised Learning

In Human-Computer Interaction

Citations in the corpus

Generating Videos with Scene Dynamics (2016, background): “Finally, this paper is related to a growing body of work that capitalizes on large amounts of unlabeled video for visual recognition tasks [18, 47, 37, 13, 24, 25, 3, 32, 26, 27, 19, 41, 42, 1].”
SoundNet: Learning Sound Representations from Unlabeled Video (2016, background): “Cross-Modal Learning and Unlabeled Video: Our approach is broadly inspired by efforts to model cross-modal relations [24, 14, 7, 26] and works that leverage large amounts of unlabeled video [25, 41, 8, 40, 39].”
Computational Baby Learning (2014, background): “For the second category, video-based approaches [19][29][1] utilized intrinsic motion cues and appearance correlations within temporal adjacent frames to augment the model training.”
Anticipating the future by watching unlabeled video (2015, methodology): “Chen and Grauman [1] propose a method to discover new human actions by only analyzing unlabeled videos., and Mobhai et al. [25] similarly discover objects.”
Track and Transfer: Watching Videos to Simulate Strong Human Supervision for Weakly-Supervised Object Detection (2016, unknown): “Existing methods learn part-based animal models [29], learn detectors from images while using video patches for regularization [20], or augment training data from videos for single-image action recognition [2].”
3D Pose from Motion for Cross-View Action Recognition via Non-linear Circulant Temporal Encoding (2014, unknown): “Another related technique, proposed by Chen and Grauman [5], uses unlabelled video data to generate examples for actions in still images.”
Anticipating Visual Representations from Unlabeled Video (2015, methodology): “Chen and Grauman [1] propose a method to discover new human actions by only analyzing unlabeled videos, and Mobahi et al. [24] similarly discover objects.”
Learning Image Representations Tied to Egomotion from Unlabeled Video (2017, background): “Temporal coherence of inferred body poses in unlabeled video is exploited for invariant recognition in [6].”