| 2024 | 2023 | 2022 | 2021 | 2020 | |||||
| 2019 | 2018 | 2017 | 2016 | 2015 | 2014 | 2013 | 2012 | 2011 | 2010 |
| 2009 | 2008 | 2007 | 2006 | 2005 | 2004 | 2003 | 2002 | 2001 | 2000 |
| 1999 | 1998 | 1997 | 1996 | 1995 | 1994 | 1993 | 1992 | 1991 | 1990 |
| 1989 | 1988 | 1987 | 1986 | 1985 | 1984 | 1983 |
- Jenkin, M. The Stereopsis of Time-Varying Images. RBCV-TR-84-3. Technical Report on Research in Biological and Computational Vision. Department of Computer Science, University of Toronto, 1984.
The goal of this thesis was to design a noncooperative algorithm for the problem of the stereopsis of time-varying imagery. The algorithm has been implemented in a comptuer system capable of interpreting a three-dimensional visual scene as presented in a sequence of stereoscopic images. The input to the system consists of a sequence of digitized stereoscopic snapshots. The output of the system is a description of the three-dimensional motoin of the objects being viewed by the system.The system integrates two problems; the problem of stereopsis and the problem of tracking objects through time. Rather than finding the intersectino of the two problems to be more difficult than solving each problem separately, it was found that by solving the two problems simultaneously, and thus incorporating the spatio-temporal context within which the scene exists, some of the hard subproblems belonging to the problems of stereopsis and temporal correspondence can be avoided.In terms of the model for temporal stereopsis presented in this thesis the induced effect and hysteresis have simple explanations as a result of the motino of the stretching simuli. It is argued that the difficulty algorithms for static stereopsis have with the induced effect and hysteresis are a result of their temporal nature.The system relies on a general smoothness assumption to assign both disparity and temporal matches. A simple model of the motion of three-dimensional points is used to guide the matching process and to identify conditional matches which violate the general smoothness assumption. A proximal rule is used to further restrict possible matches.
The system has been tested on both synthetic and real input sequences. Input sequences were chosen from three-dimensional moving light displays and "real" grey-level digitized images, using as monocular primitives interesting points detected by the Moravec operator.
- Gershon, R., Ali, Y., and Jenkin, M., An explanation system for frame-based knolwedge organized along multiple dimensions. Proc Fifth Nat. Conf. Canad. Soc. for Computational Studies of Intelligence, London, Ontario, 133-137, 1984.
This paper describes an explanation system for frame-based knowledge about events, as presented in a visual motion expert system. As such, it can be applied to representations that embody different frame organizational relationships such as is_a, part_of, instance_of, similarity and time. This may be contrasted with most current expert system which employ rule-based knowledge representations. In addition, such expert systems typically do not deal with complex spatio-temporal information and only a small number of these have any explanation capabilities. The system described in this document is capable of making inferences about frame comparisons and temporal relationships not present in the knowledge base, and provides output in both textual and pictorial formats. The graphical format is particularly useful for revealing the structure of the knowledge frames. The system has been implemented and tested on a knowledge base designed for human left ventricular peformance and examples of interaction with the system will be presented. - Jenkin, M. Applying temporal constraints to the stereopsis of time-varying imagery. Proc Fifth Nat. Conf. Canad. Soc. for Computational Studies of Intelligence, London, Ontario, 106-113, 1984.
A noncooperative algorithm is presented for the problem of the stereopsis of time-varying imagery. The algorithm integrates two problems: the problem of stereopsis and the problem of tracking objects through time. Rather than finding the intersection of the two problems to be more difficult, it was found that by solving the two problems simultaneously, and thus incorporating the spatio-temporal conttext within which a scene exists, some of the hard subproblems belonging to the problems of stereopsis and temporal correspondence could be avoided.In terms of the model for temporal stereopsis presented in this paper the induced effect and hysteresis have simple explanations as a result of the motion of the stretching stimuli. It is argued that the difficulty algorithims for static stereopsis have with the induced effect and hysteresis are a result of their temporal nature.The algorithm relies on a general smoothness assumption to assign both disparity and temporal matches. A simple model of the motion of three-dimensional points is used to guide the matching process and to identify conditional matches which violate the general smoothness assumption. A proximal rule is used to futher restrict possible matches.The algorithm has been tested on both synthetic and real input sequences. Input sequences were chosen from three-dimensional moving light displays and from "real" grey-level digitized images.
