A
Andrew Zisserman
Researcher at University of Oxford
Publications - 808
Citations - 312028
Andrew Zisserman is an academic researcher from University of Oxford. The author has contributed to research in topics: Convolutional neural network & Real image. The author has an hindex of 167, co-authored 808 publications receiving 261717 citations. Previous affiliations of Andrew Zisserman include University of Edinburgh & Microsoft.
Papers
More filters
Posted Content
Counting Out Time: Class Agnostic Video Repetition Counting in the Wild
TL;DR: This model substantially exceeds the state of the art performance on existing periodicity (PERTUBE) and repetition counting (QUVA) benchmarks and collects a new challenging dataset called Countix which captures the challenges of repetition counting in real-world videos.
Book ChapterDOI
Multiple View Geometry in Computer Vision: Epipolar Geometry and the Fundamental Matrix
Richard Hartley,Andrew Zisserman +1 more
TL;DR: The epipolar geometry is the intrinsic projective geometry between two views as discussed by the authors, which is the geometry of the intersection of the image planes with the pencil of planes having the baseline as axis (the baseline is the line joining the camera centres).
Proceedings ArticleDOI
Speech2Action: Cross-Modal Supervision for Action Recognition
TL;DR: In this paper, a BERT-based Speech2Action classifier was trained on over a thousand movie screenplays to predict action labels from transcribed speech segments and applied this model to the speech segments of a large unlabeled movie corpus (188M speech segments from 288K movies).
Book ChapterDOI
Using Projective Invariants for Constant Time Library Indexing in Model Based Vision
TL;DR: Progress is described in building a large model based vision system which uses many projectively invariant descriptors, and it is demonstrated the ease of model acquisition in the system, where models are generated directly from images.
Proceedings ArticleDOI
Finding nemo: Deformable object class modelling using curve matching
TL;DR: This work incorporates correspondence variation into the optimization, thereby extending contour-based reconstruction techniques to deformable object modelling and enables effective de-formable class reconstruction from unordered images, despite significant occlusion and the scarcity of shared 2D image features.