The Experts below are selected from a list of 2811 Experts worldwide ranked by ideXlab platform
Anton Van Den Hengel - One of the best experts on this subject based on the ideXlab platform.
-
Mid-level deep pattern mining
Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2015Co-Authors: Yao Li, Lingqiao Liu, Chunhua Shen, Anton Van Den HengelAbstract:Mid-level visual element discovery aims to find clusters of image patches that are both representative and discriminative. In this work, we study this problem from the prospective of pattern mining while relying on the recently popularized Convolutional Neural Networks (CNNs). Specifically, we find that for an image patch, activation extracted from the first fully-connected layer of a CNN have two appealing properties which enable its seamless integration with pattern mining. Patterns are then discovered from a large number of CNN activations of image patches through the well-known association rule mining. When we retrieve and visualize image patches with the same pattern (See Fig. 1), surprisingly, they are not only visually similar but also semantically consistent. We apply our approach to scene and object classification tasks, and demonstrate that our approach outperforms all previous works on mid-level visual element discovery by a sizeable margin with far fewer elements being used. Our approach also outperforms or matches recent works using CNN for these tasks. Source code of the complete system is available online.
-
mid level deep pattern mining
arXiv: Computer Vision and Pattern Recognition, 2014Co-Authors: Lingqiao Liu, Chunhua Shen, Anton Van Den HengelAbstract:Mid-level visual element discovery aims to find clusters of image patches that are both representative and discriminative. In this work, we study this problem from the prospective of pattern mining while relying on the recently popularized Convolutional Neural Networks (CNNs). Specifically, we find that for an image patch, activations extracted from the first fully-connected layer of CNNs have two appealing properties which enable its seamless integration with pattern mining. Patterns are then discovered from a large number of CNN activations of image patches through the well-known association rule mining. When we retrieve and visualize image patches with the same pattern, surprisingly, they are not only visually similar but also semantically consistent. We apply our approach to scene and object classification tasks, and demonstrate that our approach outperforms all previous works on mid-level visual element discovery by a sizeable margin with far fewer elements being used. Our approach also outperforms or matches recent works using CNN for these tasks. Source code of the complete system is available online.
Yao Li - One of the best experts on this subject based on the ideXlab platform.
-
Mid-level deep pattern mining
Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2015Co-Authors: Yao Li, Lingqiao Liu, Chunhua Shen, Anton Van Den HengelAbstract:Mid-level visual element discovery aims to find clusters of image patches that are both representative and discriminative. In this work, we study this problem from the prospective of pattern mining while relying on the recently popularized Convolutional Neural Networks (CNNs). Specifically, we find that for an image patch, activation extracted from the first fully-connected layer of a CNN have two appealing properties which enable its seamless integration with pattern mining. Patterns are then discovered from a large number of CNN activations of image patches through the well-known association rule mining. When we retrieve and visualize image patches with the same pattern (See Fig. 1), surprisingly, they are not only visually similar but also semantically consistent. We apply our approach to scene and object classification tasks, and demonstrate that our approach outperforms all previous works on mid-level visual element discovery by a sizeable margin with far fewer elements being used. Our approach also outperforms or matches recent works using CNN for these tasks. Source code of the complete system is available online.
Lingqiao Liu - One of the best experts on this subject based on the ideXlab platform.
-
Mid-level deep pattern mining
Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2015Co-Authors: Yao Li, Lingqiao Liu, Chunhua Shen, Anton Van Den HengelAbstract:Mid-level visual element discovery aims to find clusters of image patches that are both representative and discriminative. In this work, we study this problem from the prospective of pattern mining while relying on the recently popularized Convolutional Neural Networks (CNNs). Specifically, we find that for an image patch, activation extracted from the first fully-connected layer of a CNN have two appealing properties which enable its seamless integration with pattern mining. Patterns are then discovered from a large number of CNN activations of image patches through the well-known association rule mining. When we retrieve and visualize image patches with the same pattern (See Fig. 1), surprisingly, they are not only visually similar but also semantically consistent. We apply our approach to scene and object classification tasks, and demonstrate that our approach outperforms all previous works on mid-level visual element discovery by a sizeable margin with far fewer elements being used. Our approach also outperforms or matches recent works using CNN for these tasks. Source code of the complete system is available online.
-
mid level deep pattern mining
arXiv: Computer Vision and Pattern Recognition, 2014Co-Authors: Lingqiao Liu, Chunhua Shen, Anton Van Den HengelAbstract:Mid-level visual element discovery aims to find clusters of image patches that are both representative and discriminative. In this work, we study this problem from the prospective of pattern mining while relying on the recently popularized Convolutional Neural Networks (CNNs). Specifically, we find that for an image patch, activations extracted from the first fully-connected layer of CNNs have two appealing properties which enable its seamless integration with pattern mining. Patterns are then discovered from a large number of CNN activations of image patches through the well-known association rule mining. When we retrieve and visualize image patches with the same pattern, surprisingly, they are not only visually similar but also semantically consistent. We apply our approach to scene and object classification tasks, and demonstrate that our approach outperforms all previous works on mid-level visual element discovery by a sizeable margin with far fewer elements being used. Our approach also outperforms or matches recent works using CNN for these tasks. Source code of the complete system is available online.
Jacob Anglister - One of the best experts on this subject based on the ideXlab platform.
-
NMR Mapping of the IFNAR1-EC Binding Site on IFNα2 Reveals Allosteric Changes in the IFNAR2-EC Binding Site
Biochemistry, 2010Co-Authors: Sabine R. Akabayov, Zohar Biron, Jacob AnglisterAbstract:All type I interferons (IFNs) bind to a common cell-surface receptor consisting of two subunits. IFNs initiate intracellular signal transduction cascades by simultaneous interaction with the extracellular domains of its receptor subunits, IFNAR1 and IFNAR2. In this study, we mapped the surface of IFNα2 interacting with the extracellular domain of IFNAR1 (IFNAR1-EC) by following changes in or the disappearance of the 1H−15N TROSY-HSQC cross peaks of IFNα2 caused by the binding of the extracellular domain of IFNAR1 (IFNAR1-EC) to the binary complex of IFNα2 with IFNAR2-EC. The NMR study of the 89 kDa complex was conducted at pH 8 and 308 K using an 800 MHz spectrometer. IFNAR1 binding affected a total of 47 of 165 IFNα2 residues contained in two large patches on the face of the protein opposing the binding site for IFNAR2 and in a third patch located on the face containing the IFNAR2 binding site. The first two patches form the IFNAR1 binding site, and one of these matches the IFNAR1 binding site previously...
Chunhua Shen - One of the best experts on this subject based on the ideXlab platform.
-
Mid-level deep pattern mining
Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2015Co-Authors: Yao Li, Lingqiao Liu, Chunhua Shen, Anton Van Den HengelAbstract:Mid-level visual element discovery aims to find clusters of image patches that are both representative and discriminative. In this work, we study this problem from the prospective of pattern mining while relying on the recently popularized Convolutional Neural Networks (CNNs). Specifically, we find that for an image patch, activation extracted from the first fully-connected layer of a CNN have two appealing properties which enable its seamless integration with pattern mining. Patterns are then discovered from a large number of CNN activations of image patches through the well-known association rule mining. When we retrieve and visualize image patches with the same pattern (See Fig. 1), surprisingly, they are not only visually similar but also semantically consistent. We apply our approach to scene and object classification tasks, and demonstrate that our approach outperforms all previous works on mid-level visual element discovery by a sizeable margin with far fewer elements being used. Our approach also outperforms or matches recent works using CNN for these tasks. Source code of the complete system is available online.
-
mid level deep pattern mining
arXiv: Computer Vision and Pattern Recognition, 2014Co-Authors: Lingqiao Liu, Chunhua Shen, Anton Van Den HengelAbstract:Mid-level visual element discovery aims to find clusters of image patches that are both representative and discriminative. In this work, we study this problem from the prospective of pattern mining while relying on the recently popularized Convolutional Neural Networks (CNNs). Specifically, we find that for an image patch, activations extracted from the first fully-connected layer of CNNs have two appealing properties which enable its seamless integration with pattern mining. Patterns are then discovered from a large number of CNN activations of image patches through the well-known association rule mining. When we retrieve and visualize image patches with the same pattern, surprisingly, they are not only visually similar but also semantically consistent. We apply our approach to scene and object classification tasks, and demonstrate that our approach outperforms all previous works on mid-level visual element discovery by a sizeable margin with far fewer elements being used. Our approach also outperforms or matches recent works using CNN for these tasks. Source code of the complete system is available online.
