The Experts below are selected from a list of 222 Experts worldwide ranked by ideXlab platform
Yumeng Li - One of the best experts on this subject based on the ideXlab platform.
-
scale recursive network with point supervision for cRowd scene analysis
Neurocomputing, 2020Co-Authors: Zihao Dong, Ruixun Zhang, Xiuli Shao, Yumeng LiAbstract:Abstract CRowd scene analysis, and in particular its density estimation, is a challenging task due to the lack of spatial information, scale variation, and the large amount of supervised-learning parameters. In order to address these challenges, we propose a Scale-Recursive encoder–decoder Network with Point Supervision (SRN+PS). On the one hand, an encoder–decoder recurrent structure uses features between adjacent scales to tackle scale variation, and a novel loss function, called the Row Vector-based counting loss, is proposed to focus on the cRowd counting accuracy. On the other hand, we employ an additional point segmentation task in training and combine features learned from the two tasks above. The Euclidean loss, Row Vector-based counting loss, and two-label focal loss are integrated by a joint training scheme, which improves both the quality of density map estimation and the performance of cRowd counting. Finally, we propose a weakly supervised framework based on the SRN structure and the Convolutional Winner-Take-All(CWTA) module. In this framework, most parameters are obtained by unsupervised learning with the exception of a few which are tuned by supervised learning in model training. As a result, our multi-scale structure can obtain salient object sparse spatial features from unsupervised learning. Experiments on the ShanghaiTech, UCF _ CC _ 50 and UCSD datasets demonstrate the effectiveness of our proposed method.
Gilles Villard - One of the best experts on this subject based on the ideXlab platform.
-
MATRIX RANK CERTIFICATION
The Electronic Journal of Linear Algebra, 2004Co-Authors: B. David Saunders, Arne Storjohann, Gilles VillardAbstract:Randomized algorithmsare given for computing the rank of a matrix over a field of characteristic zero with conjugation operator. The matrix is treated as a black box. Only the capability to compute matrix×column-Vector and Row-Vector×matrix productsisus ed. The methods are exact, sometimes called seminumeric. They are appropriate for example for matrices with integer or rational entries. The rank algorithms are probabilistic of the Las Vegas type; the correctness of the result is guaranteed.
-
Matrix Rank Certification
2001Co-Authors: David Saunders, Arne Storjohann, Gilles VillardAbstract:Randomized algorithms are given for computing the rank of a matrix over a field of characteristic zero. The matrix is treated as a black box. Only the capability to compute matrix x column-Vector and Row-Vector x matrix products is used. The methods are exact, sometimes called seminumeric. They are appropriate for example for matrices with integer or rational entries. The rank algorithms are probabilistic of the Las Vegas type; the correctness of the result is guaranteed.
Zihao Dong - One of the best experts on this subject based on the ideXlab platform.
-
scale recursive network with point supervision for cRowd scene analysis
Neurocomputing, 2020Co-Authors: Zihao Dong, Ruixun Zhang, Xiuli Shao, Yumeng LiAbstract:Abstract CRowd scene analysis, and in particular its density estimation, is a challenging task due to the lack of spatial information, scale variation, and the large amount of supervised-learning parameters. In order to address these challenges, we propose a Scale-Recursive encoder–decoder Network with Point Supervision (SRN+PS). On the one hand, an encoder–decoder recurrent structure uses features between adjacent scales to tackle scale variation, and a novel loss function, called the Row Vector-based counting loss, is proposed to focus on the cRowd counting accuracy. On the other hand, we employ an additional point segmentation task in training and combine features learned from the two tasks above. The Euclidean loss, Row Vector-based counting loss, and two-label focal loss are integrated by a joint training scheme, which improves both the quality of density map estimation and the performance of cRowd counting. Finally, we propose a weakly supervised framework based on the SRN structure and the Convolutional Winner-Take-All(CWTA) module. In this framework, most parameters are obtained by unsupervised learning with the exception of a few which are tuned by supervised learning in model training. As a result, our multi-scale structure can obtain salient object sparse spatial features from unsupervised learning. Experiments on the ShanghaiTech, UCF _ CC _ 50 and UCSD datasets demonstrate the effectiveness of our proposed method.
Ricardo David Katz - One of the best experts on this subject based on the ideXlab platform.
-
REACHABILITY PROBLEMS FOR PRODUCTS OF MATRICES IN SEMIRINGS
International Journal of Algebra and Computation, 2006Co-Authors: Stéphane Gaubert, Ricardo David KatzAbstract:We consider the following matrix reachability problem: given $r$ square matrices with entries in a semiring, is there a product of these matrices which attains a prescribed matrix? We define similarly the Vector (resp. scalar) reachability problem, by requiring that the matrix product, acting by right multiplication on a prescribed Row Vector, gives another prescribed Row Vector (resp. when multiplied at left and right by prescribed Row and column Vectors, gives a prescribed scalar). We show that over any semiring, scalar reachability reduces to Vector reachability which is equivalent to matrix reachability, and that for any of these problems, the specialization to any $r\geq 2$ is equivalent to the specialization to $r=2$. As an application of this result and of a theorem of Krob, we show that when $r=2$, the Vector and matrix reachability problems are undecidable over the max-plus semiring $(Z\cup\{-\infty\},\max,+)$. We also show that the matrix, Vector, and scalar reachability problems are decidable over semirings whose elements are ``positive'', like the tropical semiring $(N\cup\{+\infty\},\min,+)$.
-
Reachability problems for products of matrices in semirings
2003Co-Authors: Stéphane Gaubert, Ricardo David KatzAbstract:We consider the following matrix reachability problem: given r square matrices with entries in a semiring, is there a product of these matrices which attains a prescribed matrix? We define similarly the Vector (resp. scalar) reachability problem, by requiring that the matrix product, acting by right multiplication on a prescribed Row Vector, gives another prescribed Row Vector (resp. when multiplied at left and right by prescribed Row and column Vectors, gives a prescribed scalar). We show that over any semiring, scalar reachability reduces to Vector reachability which is equivalent to matrix reachability, and that for any of these problems, the specialization to any r >= 2 is equivalent to the specialization to r=2. As an application of this result and of a theorem of Krob, we show that when r=2, the Vector and matrix reachability problems are undecidable over the max-plus semiring (Z U {-infinity},max,+). We also show that the matrix, Vector, and scalar reachability problems are decidable over semirings whose elements are «positive», like the tropical semiring (N U {+infinity},min,+).
A.j. Pritchard - One of the best experts on this subject based on the ideXlab platform.
-
New robustness results for linear systems under real perturbations
Proceedings of the 27th IEEE Conference on Decision and Control, 1Co-Authors: Diederich Hinrichsen, A.j. PritchardAbstract:Structured real perturbations of the system matrix A to A+ Delta A=BDC (with B, C given matrices) are considered. Robustness measures with respect to arbitrary stability domains C/sub g/ contained in/implied by C are introduced and characterized. These general formulas provide insight but are difficult to evaluate. Computable formulas are obtained when B is a column or C a Row Vector. >
