Learning image representations equivariant to ego-motion

By Dinesh Jayaraman, Kristen Grauman. 2015

36 citations recorded by the source corpus.

Subfields: Representation & Self-Supervised Learning

How researchers see this paper

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

In Representation & Self-Supervised Learning

Citations in the corpus

Deep multi-scale video prediction beyond mean square error (2015, methodology): “Unsupervised learning from video is also exploited in the work of Vondrick et al. (2015), where a convolutional model is trained to predict sets of future possible actions, or in (Jayaraman & Grauman, 2015) which focuses on learning a feature space equivariant to ego-motion.”
Visually Indicated Sounds (2015, background): “For example, [9] learned image features by predicting the spatial relationship between image patches, and [1, 20] by predicting the relative camera pose between frames in a video.”
Equivariance Through Parameter-Sharing (2017, background): “Deep and hierarchical models have used a variety of techniques to study or obtain representations that isolate transformations from the content “content” (Hinton et al., 2011; Jayaraman & Grauman, 2015; Lenc & Vedaldi, 2015; Agrawal et al., 2015, e.g. , ).”
Slow and Steady Feature Analysis: Higher Order Temporal Coherence in Video (2015, background): “In [14], an equivariant feature space is sought, such that the learned features respond systematically to transformations associated with distinct ego-motions, while in [2] the CNN targets the camera pose change for an input pair as regression outputs.”
Object-Centric Representation Learning from Unlabeled Videos (2016, methodology): “Baselines We compare to several existing unsupervised feature learning methods [13,14,7,12].”
Unsupervised Learning using Sequential Verification for Action Recognition (2016, background): “Several recent papers [52, 62, 64] use egomotion constraints from video to further constrain the learning.”
Shuffle and Learn: Unsupervised Learning Using Temporal Order Verification (2016, background): “Several recent papers [36, 48, 50] use egomotion constraints from video to further constrain the learning.”
Contextual Learning (2020, unknown): “(b) A less strict alternative is to only assume that all transformations are linear in s. Jayaraman & Grauman (2015) suggest to learn these transformations as an auxiliary task using the pattern depicted in Fig.”
Learning to Extract Motion from Videos in Convolutional Neural Networks (2016, background): “Note that this paper abuses of the term “invariance” for cases more accurately involving equi variance or co variance [18].”