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
Seeing wake words: Audio-visual Keyword Spotting
TL;DR: A novel convolutional architecture, KWS-Net, that uses a similarity map intermediate representation to separate the task into sequence matching, and pattern detection, to decide whether and when a word of interest is spoken by a talking face, with or without the audio.
Proceedings ArticleDOI
Shape from Shading in the Light of Mutual Illumination.
David Forsyth,Andrew Zisserman +1 more
TL;DR: It is demonstrated that for a large class of shapes the response of edge detectors will also remain unchanged and the consequences of mutual illumination effects for different theories of recovering shape from radiance are discussed.
Posted Content
Sim2real transfer learning for 3D pose estimation: motion to the rescue
Carl Doersch,Andrew Zisserman +1 more
TL;DR: This paper shows that standard neural-network approaches, which perform poorly when trained on synthetic RGB images, can perform well when the data is pre-processed to extract cues about the person's motion, notably as optical flow and the motion of 2D keypoints.
Posted Content
Deep Frank-Wolfe For Neural Network Optimization
TL;DR: The authors proposed a composite proximal framework based on the Frank-Wolfe (FW) algorithm for SVM, which computes an optimal step-size in closed-form at each time-step.
Book ChapterDOI
Projective Reconstruction of Surfaces of Revolution
Sven Utcke,Andrew Zisserman +1 more
TL;DR: In this paper, the problem of recovering the generating curve of a surface of revolution from a single uncalibrated perspective view, based solely on the object's outline and two (partly) visible cross-sections, is addressed.