The Experts below are selected from a list of 325221 Experts worldwide ranked by ideXlab platform
Philippe Cudré-mauroux - One of the best experts on this subject based on the ideXlab platform.
-
D$^2$2HistoSketch: Discriminative and Dynamic Similarity-Preserving Sketching of Streaming Histograms
IEEE Transactions on Knowledge and Data Engineering, 2019Co-Authors: Dingqi Yang, Bin Li, Laura Rettig, Philippe Cudré-maurouxAbstract:Histogram-based similarity has been widely adopted in many machine learning tasks. However, measuring Histogram similarity is a challenging task for streaming Histograms, where the elements of a Histogram are observed one after the other in an online manner. The ever-growing cardinality of Histogram elements over the data streams makes any similarity computation inefficient in that case. To tackle this problem, we propose in this paper D2HistoSketch, a similarity-preserving sketching method for streaming Histograms to efficiently approximate their Discriminative and Dynamic similarity. D2HistoSketch can fast and memory-efficiently maintain a set of compact and fixed-size sketches of streaming Histograms to approximate the similarity between Histograms. To provide high-quality similarity approximations, D2HistoSketch considers both discriminative and gradual forgetting weights for similarity measurement, and seamlessly incorporates them in the sketches. Based on both synthetic and real-world datasets, our empirical evaluation shows that our method is able to efficiently and effectively approximate the similarity between streaming Histograms while outperforming state-of-the-art sketching methods. Compared to full streaming Histograms with both discriminative and gradual forgetting weights in particular, D2HistoSketch is able to dramatically reduce the classification time (with a 7500x speedup) at the expense of a small loss in accuracy only (about 3.25 percent).
-
HistoSketch: Fast Similarity-Preserving Sketching of Streaming Histograms with Concept Drift
2017 IEEE International Conference on Data Mining (ICDM), 2017Co-Authors: Dingqi Yang, Bin Li, Laura Rettig, Philippe Cudré-maurouxAbstract:Histogram-based similarity has been widely adopted in many machine learning tasks. However, measuring Histogram similarity is a challenging task for streaming data, where the elements of a Histogram are observed in a streaming manner. First, the ever-growing cardinality of Histogram elements makes any similarity computation inefficient. Second, the concept-drift issue in the data streams also impairs the accurate assessment of the similarity. In this paper, we propose to overcome the above challenges with HistoSketch, a fast similarity-preserving sketching method for streaming Histograms with concept drift. Specifically, HistoSketch is designed to incrementally maintain a set of compact and fixed-size sketches of streaming Histograms to approximate similarity between the Histograms, with the special consideration of gradually forgetting the outdated Histogram elements. We evaluate HistoSketch on multiple classification tasks using both synthetic and real-world datasets. The results show that our method is able to efficiently approximate similarity for streaming Histograms and quickly adapt to concept drift. Compared to full streaming Histograms gradually forgetting the outdated Histogram elements, HistoSketch is able to dramatically reduce the classification time (with a 7500x speedup) with only a modest loss in accuracy (about 3.5%).
Shree K. Nayar - One of the best experts on this subject based on the ideXlab platform.
-
Use of Histograms for recognition
2020Co-Authors: Shree K. Nayar, Efstathios HadjidemetriouAbstract:Histograms have been used extensively for recognition and for retrieval of images and video from visual databases. They are efficient and have been found experimentally to be robust to certain types of image morphisms, such as viewpoint changes and object deformations. The precise effect of these image morphisms on the Histogram has not been studied. The first topic examined in this work are the transformations that preserve the Histogram of any arbitrary image they are applied to. In particular, the complete class of Histogram preserving continuous transformations is derived. Several applications of this class of transformations are shown and their significance for Histogram based image indexing is discussed. Individual Histograms of images at resolutions lower than that of the original image have also been used for image indexing. A single image Histogram, however, suffers from the inability to encode spatial image variation. Spatial variation can be incorporated into Histograms simply by taking intensity Histograms of an image at multiple resolutions to form a multiresolution Histogram. It is shown that for several classes of images the multiresolution Histogram depends on parameters and properties of image shapes and textures. Two characteristics of the multiresolution Histograms that significantly affect their indexing performance are examined. The first is the intensity resolution of the Histograms. The second is the bin width of the Histograms. The bin width can be identical for the Histograms of all image resolutions or it can depend on image resolution. Two different norms are used to compute the distance between multiresolution Histograms. For the first norm, the intensity Histograms of the various image resolutions are concatenated to form a feature vector. For the second norm, the differences between the intensity Histograms of consecutive image resolutions are concatenated to form a feature vector. The distance between the feature vectors for both norms was computed using L1. Multiresolution Histograms, like single Histograms, can be computed, stored, and matched efficiently. They also retain the robustness of the plain Histograms. The ability of multiresolution Histograms to discriminate between images of different classes is demonstrated experimentally. The experiments are performed using three databases. The first is a database of synthetic images. The second is a database of Brodatz textures [19]. The last database consists of CUReT textures [35]. Multiresolution Histograms are shown to be robust to rotations, noise, database size, and intensity resolution. The performance of the multiresolution Histogram as an image feature is compared to that of five other commonly used image features. The five image descriptors are Fourier power spectrum features, Gabor wavelet features, Daubechies wavelet packets energies, auto-cooccurrence matrices, and Markov random field parameters. The multiresolution Histogram is found to be very efficient and the most robust feature. Finally, the Shannon entropy of the multiresolution Histograms is used to reveal geometric properties of images. It is also used to select resolutions that increase the discriminability based on Histograms between different image classes.
-
Multiresolution Histograms and their use for recognition
IEEE Transactions on Pattern Analysis and Machine Intelligence, 2004Co-Authors: Efstathios Hadjidemetriou, Michael Grossberg, Shree K. NayarAbstract:The Histogram of image intensities is used extensively for recognition and for retrieval of images and video from visual databases. A single image Histogram, however, suffers from the inability to encode spatial image variation. An obvious way to extend this feature is to compute the Histograms of multiple resolutions of an image to form a multiresolution Histogram. The multiresolution Histogram shares many desirable properties with the plain Histogram, including that they are both fast to compute, space efficient, invariant to rigid motions, and robust to noise. In addition, the multiresolution Histogram directly encodes spatial information. We describe a simple yet novel matching algorithm based on the multiresolution Histogram that uses the differences between Histograms of consecutive image resolutions. We evaluate it against five widely used image features. We show that with our simple feature we achieve or exceed the performance obtained with more complicated features. Further, we show our algorithm to be the most efficient and robust.
-
Histogram Preserving Image Transformations
International Journal of Computer Vision, 2001Co-Authors: Efstathios Hadjidemetriou, Michael D. Grossberg, Shree K. NayarAbstract:Histograms are used to analyze and index images. They have been found experimentally to have low sensitivity to certain types of image morphisms, for example, viewpoint changes and object deformations. The precise effect of these image morphisms on the Histogram, however, has not been studied. In this work we derive the complete class of local transformations that preserve or scale the magnitude of the Histogram of all images. We also derive a more general class of local transformations that preserve the Histogram relative to a particular image. To achieve this, the transformations are represented as solutions to families of vector fields acting on the image. The local effect of fixed points of the fields on the Histograms is also analyzed. The analytical results are verified with several examples. We also discuss several applications and the significance of these transformations for Histogram indexing.
Efstathios Hadjidemetriou - One of the best experts on this subject based on the ideXlab platform.
-
Use of Histograms for recognition
2020Co-Authors: Shree K. Nayar, Efstathios HadjidemetriouAbstract:Histograms have been used extensively for recognition and for retrieval of images and video from visual databases. They are efficient and have been found experimentally to be robust to certain types of image morphisms, such as viewpoint changes and object deformations. The precise effect of these image morphisms on the Histogram has not been studied. The first topic examined in this work are the transformations that preserve the Histogram of any arbitrary image they are applied to. In particular, the complete class of Histogram preserving continuous transformations is derived. Several applications of this class of transformations are shown and their significance for Histogram based image indexing is discussed. Individual Histograms of images at resolutions lower than that of the original image have also been used for image indexing. A single image Histogram, however, suffers from the inability to encode spatial image variation. Spatial variation can be incorporated into Histograms simply by taking intensity Histograms of an image at multiple resolutions to form a multiresolution Histogram. It is shown that for several classes of images the multiresolution Histogram depends on parameters and properties of image shapes and textures. Two characteristics of the multiresolution Histograms that significantly affect their indexing performance are examined. The first is the intensity resolution of the Histograms. The second is the bin width of the Histograms. The bin width can be identical for the Histograms of all image resolutions or it can depend on image resolution. Two different norms are used to compute the distance between multiresolution Histograms. For the first norm, the intensity Histograms of the various image resolutions are concatenated to form a feature vector. For the second norm, the differences between the intensity Histograms of consecutive image resolutions are concatenated to form a feature vector. The distance between the feature vectors for both norms was computed using L1. Multiresolution Histograms, like single Histograms, can be computed, stored, and matched efficiently. They also retain the robustness of the plain Histograms. The ability of multiresolution Histograms to discriminate between images of different classes is demonstrated experimentally. The experiments are performed using three databases. The first is a database of synthetic images. The second is a database of Brodatz textures [19]. The last database consists of CUReT textures [35]. Multiresolution Histograms are shown to be robust to rotations, noise, database size, and intensity resolution. The performance of the multiresolution Histogram as an image feature is compared to that of five other commonly used image features. The five image descriptors are Fourier power spectrum features, Gabor wavelet features, Daubechies wavelet packets energies, auto-cooccurrence matrices, and Markov random field parameters. The multiresolution Histogram is found to be very efficient and the most robust feature. Finally, the Shannon entropy of the multiresolution Histograms is used to reveal geometric properties of images. It is also used to select resolutions that increase the discriminability based on Histograms between different image classes.
-
Multiresolution Histograms and their use for recognition
IEEE Transactions on Pattern Analysis and Machine Intelligence, 2004Co-Authors: Efstathios Hadjidemetriou, Michael Grossberg, Shree K. NayarAbstract:The Histogram of image intensities is used extensively for recognition and for retrieval of images and video from visual databases. A single image Histogram, however, suffers from the inability to encode spatial image variation. An obvious way to extend this feature is to compute the Histograms of multiple resolutions of an image to form a multiresolution Histogram. The multiresolution Histogram shares many desirable properties with the plain Histogram, including that they are both fast to compute, space efficient, invariant to rigid motions, and robust to noise. In addition, the multiresolution Histogram directly encodes spatial information. We describe a simple yet novel matching algorithm based on the multiresolution Histogram that uses the differences between Histograms of consecutive image resolutions. We evaluate it against five widely used image features. We show that with our simple feature we achieve or exceed the performance obtained with more complicated features. Further, we show our algorithm to be the most efficient and robust.
-
Histogram Preserving Image Transformations
International Journal of Computer Vision, 2001Co-Authors: Efstathios Hadjidemetriou, Michael D. Grossberg, Shree K. NayarAbstract:Histograms are used to analyze and index images. They have been found experimentally to have low sensitivity to certain types of image morphisms, for example, viewpoint changes and object deformations. The precise effect of these image morphisms on the Histogram, however, has not been studied. In this work we derive the complete class of local transformations that preserve or scale the magnitude of the Histogram of all images. We also derive a more general class of local transformations that preserve the Histogram relative to a particular image. To achieve this, the transformations are represented as solutions to families of vector fields acting on the image. The local effect of fixed points of the fields on the Histograms is also analyzed. The analytical results are verified with several examples. We also discuss several applications and the significance of these transformations for Histogram indexing.
Dingqi Yang - One of the best experts on this subject based on the ideXlab platform.
-
D$^2$2HistoSketch: Discriminative and Dynamic Similarity-Preserving Sketching of Streaming Histograms
IEEE Transactions on Knowledge and Data Engineering, 2019Co-Authors: Dingqi Yang, Bin Li, Laura Rettig, Philippe Cudré-maurouxAbstract:Histogram-based similarity has been widely adopted in many machine learning tasks. However, measuring Histogram similarity is a challenging task for streaming Histograms, where the elements of a Histogram are observed one after the other in an online manner. The ever-growing cardinality of Histogram elements over the data streams makes any similarity computation inefficient in that case. To tackle this problem, we propose in this paper D2HistoSketch, a similarity-preserving sketching method for streaming Histograms to efficiently approximate their Discriminative and Dynamic similarity. D2HistoSketch can fast and memory-efficiently maintain a set of compact and fixed-size sketches of streaming Histograms to approximate the similarity between Histograms. To provide high-quality similarity approximations, D2HistoSketch considers both discriminative and gradual forgetting weights for similarity measurement, and seamlessly incorporates them in the sketches. Based on both synthetic and real-world datasets, our empirical evaluation shows that our method is able to efficiently and effectively approximate the similarity between streaming Histograms while outperforming state-of-the-art sketching methods. Compared to full streaming Histograms with both discriminative and gradual forgetting weights in particular, D2HistoSketch is able to dramatically reduce the classification time (with a 7500x speedup) at the expense of a small loss in accuracy only (about 3.25 percent).
-
HistoSketch: Fast Similarity-Preserving Sketching of Streaming Histograms with Concept Drift
2017 IEEE International Conference on Data Mining (ICDM), 2017Co-Authors: Dingqi Yang, Bin Li, Laura Rettig, Philippe Cudré-maurouxAbstract:Histogram-based similarity has been widely adopted in many machine learning tasks. However, measuring Histogram similarity is a challenging task for streaming data, where the elements of a Histogram are observed in a streaming manner. First, the ever-growing cardinality of Histogram elements makes any similarity computation inefficient. Second, the concept-drift issue in the data streams also impairs the accurate assessment of the similarity. In this paper, we propose to overcome the above challenges with HistoSketch, a fast similarity-preserving sketching method for streaming Histograms with concept drift. Specifically, HistoSketch is designed to incrementally maintain a set of compact and fixed-size sketches of streaming Histograms to approximate similarity between the Histograms, with the special consideration of gradually forgetting the outdated Histogram elements. We evaluate HistoSketch on multiple classification tasks using both synthetic and real-world datasets. The results show that our method is able to efficiently approximate similarity for streaming Histograms and quickly adapt to concept drift. Compared to full streaming Histograms gradually forgetting the outdated Histogram elements, HistoSketch is able to dramatically reduce the classification time (with a 7500x speedup) with only a modest loss in accuracy (about 3.5%).
Bin Li - One of the best experts on this subject based on the ideXlab platform.
-
D$^2$2HistoSketch: Discriminative and Dynamic Similarity-Preserving Sketching of Streaming Histograms
IEEE Transactions on Knowledge and Data Engineering, 2019Co-Authors: Dingqi Yang, Bin Li, Laura Rettig, Philippe Cudré-maurouxAbstract:Histogram-based similarity has been widely adopted in many machine learning tasks. However, measuring Histogram similarity is a challenging task for streaming Histograms, where the elements of a Histogram are observed one after the other in an online manner. The ever-growing cardinality of Histogram elements over the data streams makes any similarity computation inefficient in that case. To tackle this problem, we propose in this paper D2HistoSketch, a similarity-preserving sketching method for streaming Histograms to efficiently approximate their Discriminative and Dynamic similarity. D2HistoSketch can fast and memory-efficiently maintain a set of compact and fixed-size sketches of streaming Histograms to approximate the similarity between Histograms. To provide high-quality similarity approximations, D2HistoSketch considers both discriminative and gradual forgetting weights for similarity measurement, and seamlessly incorporates them in the sketches. Based on both synthetic and real-world datasets, our empirical evaluation shows that our method is able to efficiently and effectively approximate the similarity between streaming Histograms while outperforming state-of-the-art sketching methods. Compared to full streaming Histograms with both discriminative and gradual forgetting weights in particular, D2HistoSketch is able to dramatically reduce the classification time (with a 7500x speedup) at the expense of a small loss in accuracy only (about 3.25 percent).
-
HistoSketch: Fast Similarity-Preserving Sketching of Streaming Histograms with Concept Drift
2017 IEEE International Conference on Data Mining (ICDM), 2017Co-Authors: Dingqi Yang, Bin Li, Laura Rettig, Philippe Cudré-maurouxAbstract:Histogram-based similarity has been widely adopted in many machine learning tasks. However, measuring Histogram similarity is a challenging task for streaming data, where the elements of a Histogram are observed in a streaming manner. First, the ever-growing cardinality of Histogram elements makes any similarity computation inefficient. Second, the concept-drift issue in the data streams also impairs the accurate assessment of the similarity. In this paper, we propose to overcome the above challenges with HistoSketch, a fast similarity-preserving sketching method for streaming Histograms with concept drift. Specifically, HistoSketch is designed to incrementally maintain a set of compact and fixed-size sketches of streaming Histograms to approximate similarity between the Histograms, with the special consideration of gradually forgetting the outdated Histogram elements. We evaluate HistoSketch on multiple classification tasks using both synthetic and real-world datasets. The results show that our method is able to efficiently approximate similarity for streaming Histograms and quickly adapt to concept drift. Compared to full streaming Histograms gradually forgetting the outdated Histogram elements, HistoSketch is able to dramatically reduce the classification time (with a 7500x speedup) with only a modest loss in accuracy (about 3.5%).
