Sciweavers

376 search results - page 10 / 76
» Salient Frame Extraction for Structure from Motion
Sort
View
ICIP
2005
IEEE
14 years 9 months ago
Spatio-temporal attention model for video content analysis
This paper presents a new model of human attention that allows salient areas to be extracted from video frames. As automatic understanding of video semantic content is still far fr...
Mickael Guironnet, Nathalie Guyader, Denis Pelleri...
CORR
2002
Springer
160views Education» more  CORR 2002»
13 years 7 months ago
Structure from Motion: Theoretical Foundations of a Novel Approach Using Custom Built Invariants
We rephrase the problem of 3D reconstruction from images in terms of intersections of projections of orbits of custom built Lie groups actions. We then use an algorithmic method b...
Pierre-Louis Bazin, Mireille Boutin
PCM
2001
Springer
195views Multimedia» more  PCM 2001»
13 years 12 months ago
A Motion Activity Descriptor and Its Extraction in Compressed Domain
A novel motion activity descriptor and its extraction from a compressed MPEG (MPEG-1/2) video stream are presented. The descriptor consists of two parts, a temporal descriptor and ...
Xinding Sun, Ajay Divakaran, B. S. Manjunath
ICIP
2001
IEEE
14 years 9 months ago
Robust estimation of depth and motion using stochastic approximation
The problemof structurefrom motion (SFM)is to extract the three-dimensionalmodel of a moving scene from a sequence of images. Though two images are sufficient to produce a 3D reco...
Rama Chellappa, Amit K. Roy Chowdhury
SCVMA
2004
Springer
14 years 24 days ago
Structure from Periodic Motion
Abstract. We show how to exploit temporal periodicity of moving objects to perform 3D reconstruction. The collection of period-separated frames serve as a surrogate for multiple ri...
Serge J. Belongie, Josh Wills