The Experts below are selected from a list of 93 Experts worldwide ranked by ideXlab platform
Naoto Tanji - One of the best experts on this subject based on the ideXlab platform.
-
Gauge ambiguity of the quark spectrum in the Color Glass Condensate
Nuclear Physics A, 2019Co-Authors: François Gelis, Naoto TanjiAbstract:Abstract In the Color Glass Condensate, the inclusive spectrum of produced quarks in a heavy ion collision is obtained as the Fourier transform of a 2-fermion correlation function. Due to its non-locality, the two points of this function must be linked by a Wilson line in order to have a gauge invariant result, but when the quark spectrum is evaluated in a background that has a non-zero chromo-magnetic field, this procedure suffers from an ambiguity related to the choice of the Contour defining the Wilson line. In this paper, we use an analytically tractable toy model of the background field in order to study this Contour dependence. We show that for a Straight Contour, unphysical contributions to the spectrum in p ⊥ − 2 and p ⊥ − 3 cancel, leading to a spectrum with a tail in p ⊥ − 4 . If the Contour defining the Wilson line deviates from a Straight line, the path dependence is at most of order p ⊥ − 5 if its curvature is bounded, and of order p ⊥ − 4 otherwise. When the Contour is forced to go through a fixed point, the path dependence is even larger, of order p ⊥ − 2 .
Cordelia Schmid - One of the best experts on this subject based on the ideXlab platform.
-
Groups of adjacent Contour segments for object detection
2008Co-Authors: Vittorio Ferrari, Cordelia Schmid, Loic Fevrier, Frederic JurieAbstract:We present a family of scale-invariant local shape features formed by chains of k connected, roughly Straight Contour segments (kAS), and their use for object class detection. kAS are able to cleanly encode pure fragments of an object boundary, without including nearby clutter. Moreover, they oer an attractive compromise between information content and repeatability, and encompass a wide variety of local shape structures.
-
Groups of adjacent Contour segments for object detection
IEEE Transactions on Pattern Analysis and Machine Intelligence, 2008Co-Authors: Vittorio Ferrari, Frederic Jurie, Laureline Février, Cordelia SchmidAbstract:We present a family of scale-invariant local shape features formed by chains of k connected, roughly Straight Contour segments (kAS), and their use for object class detection. kAS are able to cleanly encode pure fragments of an object boundary, without including nearby clutter. Moreover, they offer an attractive compromise between information content and repeatability, and encompass a wide variety of local shape structures. We also define a translation and scale invariant descriptor encoding the geometric configuration of the segments within a kAS, making kAS easy to reuse in other frameworks, for example as a replacement or addition to interest points. Software for detecting and describing kAS is released on lear.inrialpes.fr/software. We demonstrate the high performance of kAS within a simple but powerful sliding-window object detection scheme. Through extensive evaluations, involving eight diverse object classes and more than 1400 images, we 1) study the evolution of performance as the degree of feature complexity k varies and determine the best degree; 2) show that kAS substantially outperform interest points for detecting shape-based classes; 3) compare our object detector to the recent, state-of-the-art system by Dalal and Triggs [4].
-
Groups of Adjacent Contour Segments for Object Detection
IEEE transactions on pattern analysis and machine intelligence, 2008Co-Authors: Vittorio Ferrari, Frederic Jurie, Laureline Février, Cordelia SchmidAbstract:We present a family of scale-invariant local shape features formed by chains of k connected roughly Straight Contour segments (kAS), and their use for object class detection. kAS are able to cleanly encode pure fragments of an object boundary without including nearby clutter. Moreover, they offer an attractive compromise between information content and repeatability and encompass a wide variety of local shape structures. We also define a translation and scale invariant descriptor encoding the geometric configuration of the segments within a kAS, making kAS easy to reuse in other frameworks, for example, as a replacement or addition to interest points (IPs). Software for detecting and describing kAS is released at http://lear.inrialpes.fr/software. We demonstrate the high performance of kAS within a simple but powerful sliding-window object detection scheme. Through extensive evaluations, involving eight diverse object classes and more than 1,400 images, we (1) study the evolution of performance as the degree of feature complexity k varies and determine the best degree, (2) show that kAS substantially outperform IPs for detecting shape-based classes, and (3) compare our object detector to the recent state-of-the-art system by Dalal and Triggs (2005).
-
Groups of Adjacent Contour Segments for Object Detection
2006Co-Authors: Vittorio Ferrari, Frederic Jurie, Loic Fevrier, Cordelia SchmidAbstract:We present a family of scale-invariant local shape features formed by chains of k connected, roughly Straight Contour segments (kAS), and their use for object class detection. kAS are able to cleanly encode pure fragments of an object boundary, without including nearby clutter. Moreover, they offer an attractive compromise between information content and repeatability, and encompass a wide variety of local shape structures. We also define a translation and scale invariant descriptor encoding the geometric configuration of the segments within a kAS, making $k$AS easy to reuse in other frameworks, for example as a replacement or addition to interest points. We demonstrate the high performance of kAS within a simple but powerful sliding-window object detection scheme. Through extensive evaluations, involving eight diverse object classes and more than 1400 images, we 1) study the evolution of performance as the degree of feature complexity k varies and determine the best degree; 2) show that kAS substantially outperform interest points for detecting shape-based classes; 3) compare our object detector to the recent, state-of-the-art system by Dalal and Triggs.
François Gelis - One of the best experts on this subject based on the ideXlab platform.
-
Gauge ambiguity of the quark spectrum in the Color Glass Condensate
Nuclear Physics A, 2019Co-Authors: François Gelis, Naoto TanjiAbstract:Abstract In the Color Glass Condensate, the inclusive spectrum of produced quarks in a heavy ion collision is obtained as the Fourier transform of a 2-fermion correlation function. Due to its non-locality, the two points of this function must be linked by a Wilson line in order to have a gauge invariant result, but when the quark spectrum is evaluated in a background that has a non-zero chromo-magnetic field, this procedure suffers from an ambiguity related to the choice of the Contour defining the Wilson line. In this paper, we use an analytically tractable toy model of the background field in order to study this Contour dependence. We show that for a Straight Contour, unphysical contributions to the spectrum in p ⊥ − 2 and p ⊥ − 3 cancel, leading to a spectrum with a tail in p ⊥ − 4 . If the Contour defining the Wilson line deviates from a Straight line, the path dependence is at most of order p ⊥ − 5 if its curvature is bounded, and of order p ⊥ − 4 otherwise. When the Contour is forced to go through a fixed point, the path dependence is even larger, of order p ⊥ − 2 .
Vittorio Ferrari - One of the best experts on this subject based on the ideXlab platform.
-
Groups of adjacent Contour segments for object detection
2008Co-Authors: Vittorio Ferrari, Cordelia Schmid, Loic Fevrier, Frederic JurieAbstract:We present a family of scale-invariant local shape features formed by chains of k connected, roughly Straight Contour segments (kAS), and their use for object class detection. kAS are able to cleanly encode pure fragments of an object boundary, without including nearby clutter. Moreover, they oer an attractive compromise between information content and repeatability, and encompass a wide variety of local shape structures.
-
Groups of adjacent Contour segments for object detection
IEEE Transactions on Pattern Analysis and Machine Intelligence, 2008Co-Authors: Vittorio Ferrari, Frederic Jurie, Laureline Février, Cordelia SchmidAbstract:We present a family of scale-invariant local shape features formed by chains of k connected, roughly Straight Contour segments (kAS), and their use for object class detection. kAS are able to cleanly encode pure fragments of an object boundary, without including nearby clutter. Moreover, they offer an attractive compromise between information content and repeatability, and encompass a wide variety of local shape structures. We also define a translation and scale invariant descriptor encoding the geometric configuration of the segments within a kAS, making kAS easy to reuse in other frameworks, for example as a replacement or addition to interest points. Software for detecting and describing kAS is released on lear.inrialpes.fr/software. We demonstrate the high performance of kAS within a simple but powerful sliding-window object detection scheme. Through extensive evaluations, involving eight diverse object classes and more than 1400 images, we 1) study the evolution of performance as the degree of feature complexity k varies and determine the best degree; 2) show that kAS substantially outperform interest points for detecting shape-based classes; 3) compare our object detector to the recent, state-of-the-art system by Dalal and Triggs [4].
-
Groups of Adjacent Contour Segments for Object Detection
IEEE transactions on pattern analysis and machine intelligence, 2008Co-Authors: Vittorio Ferrari, Frederic Jurie, Laureline Février, Cordelia SchmidAbstract:We present a family of scale-invariant local shape features formed by chains of k connected roughly Straight Contour segments (kAS), and their use for object class detection. kAS are able to cleanly encode pure fragments of an object boundary without including nearby clutter. Moreover, they offer an attractive compromise between information content and repeatability and encompass a wide variety of local shape structures. We also define a translation and scale invariant descriptor encoding the geometric configuration of the segments within a kAS, making kAS easy to reuse in other frameworks, for example, as a replacement or addition to interest points (IPs). Software for detecting and describing kAS is released at http://lear.inrialpes.fr/software. We demonstrate the high performance of kAS within a simple but powerful sliding-window object detection scheme. Through extensive evaluations, involving eight diverse object classes and more than 1,400 images, we (1) study the evolution of performance as the degree of feature complexity k varies and determine the best degree, (2) show that kAS substantially outperform IPs for detecting shape-based classes, and (3) compare our object detector to the recent state-of-the-art system by Dalal and Triggs (2005).
-
Groups of Adjacent Contour Segments for Object Detection
2006Co-Authors: Vittorio Ferrari, Frederic Jurie, Loic Fevrier, Cordelia SchmidAbstract:We present a family of scale-invariant local shape features formed by chains of k connected, roughly Straight Contour segments (kAS), and their use for object class detection. kAS are able to cleanly encode pure fragments of an object boundary, without including nearby clutter. Moreover, they offer an attractive compromise between information content and repeatability, and encompass a wide variety of local shape structures. We also define a translation and scale invariant descriptor encoding the geometric configuration of the segments within a kAS, making $k$AS easy to reuse in other frameworks, for example as a replacement or addition to interest points. We demonstrate the high performance of kAS within a simple but powerful sliding-window object detection scheme. Through extensive evaluations, involving eight diverse object classes and more than 1400 images, we 1) study the evolution of performance as the degree of feature complexity k varies and determine the best degree; 2) show that kAS substantially outperform interest points for detecting shape-based classes; 3) compare our object detector to the recent, state-of-the-art system by Dalal and Triggs.
Frederic Jurie - One of the best experts on this subject based on the ideXlab platform.
-
Groups of adjacent Contour segments for object detection
2008Co-Authors: Vittorio Ferrari, Cordelia Schmid, Loic Fevrier, Frederic JurieAbstract:We present a family of scale-invariant local shape features formed by chains of k connected, roughly Straight Contour segments (kAS), and their use for object class detection. kAS are able to cleanly encode pure fragments of an object boundary, without including nearby clutter. Moreover, they oer an attractive compromise between information content and repeatability, and encompass a wide variety of local shape structures.
-
Groups of adjacent Contour segments for object detection
IEEE Transactions on Pattern Analysis and Machine Intelligence, 2008Co-Authors: Vittorio Ferrari, Frederic Jurie, Laureline Février, Cordelia SchmidAbstract:We present a family of scale-invariant local shape features formed by chains of k connected, roughly Straight Contour segments (kAS), and their use for object class detection. kAS are able to cleanly encode pure fragments of an object boundary, without including nearby clutter. Moreover, they offer an attractive compromise between information content and repeatability, and encompass a wide variety of local shape structures. We also define a translation and scale invariant descriptor encoding the geometric configuration of the segments within a kAS, making kAS easy to reuse in other frameworks, for example as a replacement or addition to interest points. Software for detecting and describing kAS is released on lear.inrialpes.fr/software. We demonstrate the high performance of kAS within a simple but powerful sliding-window object detection scheme. Through extensive evaluations, involving eight diverse object classes and more than 1400 images, we 1) study the evolution of performance as the degree of feature complexity k varies and determine the best degree; 2) show that kAS substantially outperform interest points for detecting shape-based classes; 3) compare our object detector to the recent, state-of-the-art system by Dalal and Triggs [4].
-
Groups of Adjacent Contour Segments for Object Detection
IEEE transactions on pattern analysis and machine intelligence, 2008Co-Authors: Vittorio Ferrari, Frederic Jurie, Laureline Février, Cordelia SchmidAbstract:We present a family of scale-invariant local shape features formed by chains of k connected roughly Straight Contour segments (kAS), and their use for object class detection. kAS are able to cleanly encode pure fragments of an object boundary without including nearby clutter. Moreover, they offer an attractive compromise between information content and repeatability and encompass a wide variety of local shape structures. We also define a translation and scale invariant descriptor encoding the geometric configuration of the segments within a kAS, making kAS easy to reuse in other frameworks, for example, as a replacement or addition to interest points (IPs). Software for detecting and describing kAS is released at http://lear.inrialpes.fr/software. We demonstrate the high performance of kAS within a simple but powerful sliding-window object detection scheme. Through extensive evaluations, involving eight diverse object classes and more than 1,400 images, we (1) study the evolution of performance as the degree of feature complexity k varies and determine the best degree, (2) show that kAS substantially outperform IPs for detecting shape-based classes, and (3) compare our object detector to the recent state-of-the-art system by Dalal and Triggs (2005).
-
Groups of Adjacent Contour Segments for Object Detection
2006Co-Authors: Vittorio Ferrari, Frederic Jurie, Loic Fevrier, Cordelia SchmidAbstract:We present a family of scale-invariant local shape features formed by chains of k connected, roughly Straight Contour segments (kAS), and their use for object class detection. kAS are able to cleanly encode pure fragments of an object boundary, without including nearby clutter. Moreover, they offer an attractive compromise between information content and repeatability, and encompass a wide variety of local shape structures. We also define a translation and scale invariant descriptor encoding the geometric configuration of the segments within a kAS, making $k$AS easy to reuse in other frameworks, for example as a replacement or addition to interest points. We demonstrate the high performance of kAS within a simple but powerful sliding-window object detection scheme. Through extensive evaluations, involving eight diverse object classes and more than 1400 images, we 1) study the evolution of performance as the degree of feature complexity k varies and determine the best degree; 2) show that kAS substantially outperform interest points for detecting shape-based classes; 3) compare our object detector to the recent, state-of-the-art system by Dalal and Triggs.