S. Ullman and R. Basri. Recognition by Linear Combination of Models. IEEE Trans. Pattern Analysis and Machine Intelligence, 13:992--1006, 1991.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....of existing motion data is employed for the onthe fly synthesis [RCB98] WH97] 3. LINEAR COMBINATIONS OF STATIONARY OBJECTS The basic idea of the vector space representation by a linear combination of stationary objects can be described as follows, first proposed by Ulman and Basri [UB91] and followed up for 2D images [BP96] VP97] and 3D geometries [BV99] She00] It is based on a data set of stationary objects in a same class. All of these exemplar objects are assumed in full correspondence, which can be done using techniques based on optic flow algorithms [BP96] BV99] VP97] ....

....The vectors X i comprise the basis of a linear vector space. The method parametrizes a continuous class of objects and the weight vector w = w 1 , wm ) characterizes each object of this class in a compact way. Object transformations, like view point changes in 2D images of an object [UB91] or attribute manipulation in 3D faces [BV99] can be expressed in terms of changes in the weight vector. 4. LINEAR COMBINATIONS OF MOTIONS Motion can be described by a set of motion curves each giving the value of one of the model s parameters as a function of time, e.g. joint angles over time ....

S. Ulman and R. Basri. Recognition by linear combination of models. IEEE Transactions on Pattern Recognition and Machine Intelligence, 13:992 -- 1006, 1991.

....images based methods when multiple images per person are available, 2) hybrid methods when multiple training images are available during training but only one database image per person is available during recognition, and 3) single image shape based methods when no training is carried out. We have [8, 2, 31, 10] in the first type, and [32, 28, 17, 7] in the second type. Up to now, the second type of approach is the most popular one. The third approach does not seem to have received much attention. 2.3.1 Multi image based approaches Among the first class of approaches, one of the earliest is by Beymer ....

S. Ullman and R. Basri, "Recognition by Linear Combinations of Models," In IEEE Trans. on PAMI, Vol. 13, pp. 992-1006, 1991.

....perspective projection by calculating the direction vector to the vanishing point. require at least three 2D views of a 3D object inorder to reconstruct the object [50, 104] At least two views (orthographic) are required for constructing the affine structure which can be used for recognition [105, 80]. When the 3D objects are known to be symmetric, the symmetries can be exploited to reduce the number of views required. Specifically, given a single 2D projection (orthographic) of a mirror symmetric 3D object which is a non accidental view (i.e. the projection is not along the mirror plane of ....

S. Ullman and R. Basri. Recognition by linear combinations of models. IEEE Trans. on Pattern Analysis and Machine Intelligence, 13:992--1006, 1991.

....2D transformations, 3D rotations, as well as illumination changes, usually require multiple example views during learning. Mathematically, the space of the orthographic views of one object is spanned by a single view for affine 2D transformations and by two or more for affine 3D transformations [64, 47]. The basic facts summarized earlier, together with the above computational considerations, lead to a Standard Model, likely to represent the simplest class of models reflecting the known anatomical and biological constraints. The model reflects the general organization of visual cortex in a ....

S. Ullman and R. Basri. Recognition by linear combinations of models. IEEE Trans. Pattern Anal. Machine Intell., 13:992--1006, 1991.

....they avoid many of the computational difficulties of 3D model based recognition and make it feasible to apply techniques of machine learning to the problem of learning the appearance of 3D objects from 2D samples. Two major existing approaches to this problem are based on view interpolation[38, 28] and traditional classification methods from pattern recognition. In particular, the view interpolation approach has been very influential in formulating testable psychophysical and neurophysiological hypotheses. However, both approaches have some limitations, foremost that they do not explain how ....

....1 , B r and the corresponding viewing parameters V 1 , V r , find an f M such that f M (V i ) B i is minimized (for some norm ) This approach is illustrated schematically in Figure 1. This approach can be motivated by the observation of Ullman and Basri [38] for objects represented as ordered collections of point features under 3D rigid body transformations and orthographic projection. In Figure 1: Schematic illustration of recognition of 3D Objects by view interpolation. A number of training views are assumed for each object. To classify an unknown ....

[Article contains additional citation context not shown here]

Shimon Ullman and Ronen Basri. Recognition by Linear Combinations of Models. A.I. Memo No. 1152, MIT Artificial Intelligence Laboratory, Cambridge, MA, USA, 1989.

....object model that best matches the observed image. A common approach to recognition uses features (such as points or edges) to represent objects. An object is recognized in this approach if there exists a viewpoint from which the model features coincide with the corresponding image features, e.g. [4, 7, 9]. Since images often are noisy and models occasionally are imperfect, it is rarely the case that a model aligns perfectly with the image. Moreover, in problems such as classification and recognition of non rigid objects, the agreement between model and image is even less predictable. Systems ....

....rigid case because the quadratic constraints imposed in the rigid case are not taken into account, enabling the construction of a closed form solution. At least six points are required to find an affine solution under perspective projection [4] and four are required under orthographic projection [9]. The affine measure bounds the rigid measure from below. The rigid measure, however, is not bounded from above, and so the actual rigid measure may sometimes be significantly larger than the computed affine measure. A second approach to comparing models to images, often called alignment, ....

[Article contains additional citation context not shown here]

Ullman, S. and Basri, R. (1991). Recognition by linear combinations of models. IEEE Transactions on Pattern Analysis and Machine Intelligence, 13:992--1006.

No context found.

S. Ullman and R. Basri. Recognition by Linear Combination of Models. IEEE Trans. Pattern Analysis and Machine Intelligence, 13:992--1006, 1991.

No context found.

S. Ullman, R. Basri, Recognition by linear combination of models, IEEE Transactions on Pattern Analysis and Machine Intelligence 13 (1991) 992 -- 1006.

No context found.

S. Ullman and R. Basri. Recognition by linear combinations of models. IEEE Pat. Anal. Mach. Intell., 13(10):992--1006, 1991.

No context found.

S. Ullman and R. Basri. Recognition by linear combinations of models. IEEE Transactions on Pattern Analysis and Machine Intelligence, 13(10):992-1006, October 1991.

No context found.

S. Ullman and R. Basri, "Recognition by linear combinations of models," IEEE Trans. Pattern Anal. Machine Intell., vol. 13, pp. 992--1006, Oct. 1991.

No context found.

S. Ullman and R. Basri. Recognition by linear combinations of models. IEEE Transactions on Pattern Analysis and Machine Intelligence, 13(10):992--1005, 1991.

No context found.

S. Ullman and R. Basri. Recognition by a linear combination of models. A.I. Memo 1152, MIT, Aug. 1989.

No context found.

S. Ullman and R. Basri. Recognition by a linear combination of models. IEEE Trans. Pattern Anal. Mach. Intelligence, 13:992--1006, 1991.

No context found.

S. Ullman and R. Basri, \Recognition by linear combinations of models", ieee Transactions on Pattern Analysis and Machine Intelligence, vol. 13, no. 10, pp. 992-1006, 1991.

No context found.

S. Ullman & R. Basri. Recognition by linear combinations of models. IEEE Trans. PAMI, 13(10):992--1106, 1991.

No context found.

S. Ullman and R. Basri. Recognition by linear combinations of models. 13(10):992--1005, 1991.

No context found.

S. Ullman and R. Basri. Recognition by linear combinations of models. IEEE Pat. Anal. Mach. Intell., 13(10):992--1006, 1991.

No context found.

S. Ullman and R. Basri, "Recognition by linear combinations of models," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 13, no. 10, pp. 992--1005, Oct. 1991.

No context found.

# S. Ullman and R. Basri, "Recognition by a Linear Combination of Models," IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 13, pp. 992--1,006, 1991.

No context found.

Shimon Ullman and Ronen Basri. Recognition by linear combinations of models. IEEE Transactions 13(10):992--1005, Oct. 1991.

No context found.

S. Ullman and R. Basri, "Recognition by linear combination of models," IEEE Pattern Anal. Machine Intell., vol. 13, pp. 992--1006, October 1991.

No context found.

S. Ullman and R. Basri, Recognition by Linear Combinations of Models, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 13, No. 10, pp. 992-1006, 1991.

No context found.

Shimon Ullman and Ronen Basri. Recognition by Linear Combinations of Models. IEEE Transaction on Pattern Analysis and Machine Intelligence, 13:991--1006, October 1991.

No context found.

S. Ullman and R. Basri. Recognition by linear combinations of models. IEEE Transactions on Pattern Analysis and Machine Intelligence, 13:992--1006, 1991.

First 50 documents  Next 50

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC