Examplebased Learning for SingleImage Superresolution


 Letitia Berry
 1 years ago
 Views:
Transcription
1 Examplebased Learning for SingleImage Superresolution Kwang In Kim 1 and Younghee Kwon 2 1 MaxPlanckInstitute für biologische Kybernetik, Spemannstr. 38, D Tübingen, Germany 2 Korea Advanced Institute of Science and Technology, Kusongdong, YusongKu, Taejon, Korea Abstract. This paper proposes a regressionbased method for singleimage superresolution. Kernel ridge regression (KRR) is used to estimate the highfrequency details of the underlying highresolution image. A sparse solution of KRR is found by combining the ideas of kernel matching pursuit and gradient descent, which allows timecomplexity to be kept to a moderate level. To resolve the problem of ringing artifacts occurring due to the regularization effect, the regression results are postprocessed using a prior model of a generic image class. Experimental results demonstrate the effectiveness of the proposed method. 1 Introduction Singleimage superresolution refers to the task of constructing a highresolution enlargement of a given lowresolution image. This problem is inherently illposed as there are generally multiple highresolution images that can produce the same lowresolution image. Accordingly, prior information is required to approach this problem. Often, this prior information is available either in the explicit form of an energy functional defined on the image class [9, 10], or in the implicit form of example images leading to examplebased superresolution [1 3, 5]. Previous examplebased superresolution algorithms can be characterized as nearest neighbor (NN)based estimations [1 3] : during the training phase, pairs of lowresolution and the corresponding highresolution image patches (subwindows of images) are collected. Then, in the superresolution phase, each patch of the given lowresolution image is compared to the stored lowresolution patches, and the highresolution patch corresponding to the nearest lowresolution patch is selected as the output. For instance, Freeman et al. [2] posed the image superresolution as the problem of estimating missing highfrequency details by interpolating the input lowresolution image into the desired scale (which results in a blurred image). Then, the superresolution was performed by the NNbased estimation of highfrequency patches based on the corresponding patches of input lowfrequency image. Although this method (and also other NNbased methods) has already shown an impressive performance, there is still room for improvement if one views the
2 2 image superresolution as a regression problem, i.e., finding a map f from the space of lowresolution image patches X to the space of target highresolution patches Y. It is well known in the machine learning community that NNbased estimation suffers from overfitting where one obtains a function which explains the training data perfectly yet cannot be generalized to unknown data. In the superresolution, this can result in noisy reconstructions at complex image regions (cf. Sect. 3). Accordingly, it is reasonable to expect that NNbased methods can be improved by adopting learning algorithms with regularization capability to avoid overfitting. Based on the framework of Freeman et al. [2], Kim et al. posed the problem of estimating the highfrequency details as a regression problem which is then resolved by support vector regression (SVR) [6]. Meanwhile, Ni and Nguyen utilized SVR in the frequency domain and posed the superresolution as a kernel learning problem [7]. While SVR produced a significant improvement over existing examplebased methods, it has several drawbacks in building a practical system: 1. As a regularization framework, SVR tends to smooth the sharp edges and produce an oscillation along the major edges. This might lead to low reconstruction error on average, but is visually implausible; 2. SVR results in a dense solution, i.e., the regression function is expanded in the whole set of training data points and accordingly is computationally demanding both in training and in testing. 3 The current work extends the framework of Kim et al. [6]. A kernel ridge regression (KRR) is utilized for the regression. Due to the observed optimality of ɛ at (nearly) 0 for SVR in our previous study, the only difference between SVR and KRR in the proposed setting is their loss functions (L 1  and L 2  loss, respectively). The L 2 loss adopted by KRR is differentiable and facilitates gradientbased optimization. To reduce the time complexity of KRR, a sparse basis is found by combining the idea of the kernel matching pursuit (KMP) [11] and gradient descent such that the time complexity and the quality of superresolution can be traded. As the regularizer of KRR is the same as that of SVR, the problem of oscillation along the major edges still remains. This is resolved by exploiting a prior over image structure proposed by Tappen et al. [9]. 2 Regressionbased Image Superresolution Base System. Adopting the framework of Freeman et al. [2], for the superresolution of a given image, we estimate the corresponding missing highfrequency details based on its interpolation into the desired scale, which in this work is obtained by the bicubic interpolation. Furthermore, based on the conditional independence assumption of high and lowfrequency components given midfrequency components of an image [2], the estimation of highfrequency components (Y ) is performed based on the Laplacian of the bicubic interpolation (X). The Y is then added to the bicubic to produce the superresolved image Z. 3 In our simulation, the optimum value of ɛ for the ɛinsensitive loss function of SVR was close to zero.
3 3 To retain the complexity of the resulting regression problem at a moderate level, a patchbased approach is taken where the estimation of the values of Y at specific locations N N (Y (x, y)) is performed based on only the values of X at corresponding locations N M (X(x, y)), where N G (S(x, y)) represents a Gsized square window (patch) centered at the location (x, y) of the image S. Then, during the superresolution, X is scanned with a small window (of size M) to produce a patchvalued regression result (of size N) for each pixel. This results in a set of candidate pixels for each location of Z (as the patches are overlapping with their neighbors), which are then combined to make the final estimation (details will be provided later). The training images for the regressor are obtained by blurring and subsampling (by bicubic resampling) a set of highresolution images to constitute a set of low and highresolution image pairs. The training image patch pairs are randomly sampled therein. To increase the efficiency of the training set, the data are contrastnormalized ([2]): during the construction of the training set both the input image patch and corresponding desired patches are normalized by dividing them by the L 1 norm of the input patch. For an unseen image patch, the input is again normalized before the regression and the corresponding output is inverse normalized. For a given set of training data points {(x 1, y 1 ),..., (x l, y l )} IR M IR N, we minimize the following regularized cost functional O({f 1,..., f N }) = ( 1 (f i (x j ) y i 2 j) ) 2 λ f i 2 H, (1) i=1,...,n j=1,...,l where y j = [yj 1,..., yn j ] and H is a reproducing kernel Hilbert space (RKHS). Due to the reproducing property, the minimizer of above functional is expanded in kernel functions: f i ( ) = a i jk(x j, ), for i = 1,..., N (2) j=1,...,l where k is the generating kernel for H which, we choose as a Gaussian kernel (k(x, y) = exp ( x y 2 /σ k ) ). Equation (1) is the sum of individual convex cost functionals for each scalarvalued regressor and can be minimized separately. However, by tying the regularization parameter λ and the kernel k we can reduce the time complexity of training and testing down to the case of scalarvalued regression, as in this case the kernel matrix can be shared: plugging (2) into (1) and noting the convexity of (1) yields A = (K + λi) 1 Y, (3) where Y = [y 1,..., y l ] and the ith column of A constitutes the coefficient vector a i = [a i 1,..., a i l ] for the ith regressor. Sparse Solution. As evident from (2) and (3), the training and testing time of KRR is O(l 3 ) and O(M l), respectively, which becomes prohibitive even for a relatively small number of training data points (e.g., l > 10, 000). One
4 4 way of reducing the time complexity is to trade it off with the optimality of the solution by finding the minimizer of (1) only within the span of a basis set {k(b 1, ),..., k(b lb, )} (l b l): f i ( ) = a i jk(b j, ), for i = 1,..., N. (4) j=1,...,l b In this case, the solution is obtained by A = (K bx K bx + λk bb ) 1 K bx Y, (5) where [K bx(i,j) ] lb,l = k(b i, x j ) and [K bb(i,j) ] lb,l b = k(b i, b j ), and accordingly the time complexity reduces to O(M l b ) for testing. For a given fixed basis points B = {b 1,..., b lb }, the time complexity of computing the coefficient matrix A is O(lb 3 + l l b M). In general, the total training time depends on the method of finding B. In KMP [11, 4], the basis points are selected from the training data points in an incremental way: for given n 1 basis points, the nth basis is chosen such that the cost functional (1) is minimized when the A is optimized accordingly. The exact implementation of KMP costs O(l 2 )time for each step. Another possibility is to note the differentiability of the cost functional (4) which leads to gradientbased optimization to construct B. Assuming that the evaluation of the derivative of k with respect to a basis vector takes O(M)time, the evaluation of derivative of (1) with respect to B and corresponding coefficient matrix A takes O(M l l b + l lb 2 )time. Because of the increased flexibility, in general, gradientbased methods can lead to a better optimization of the cost functional (1) than selection methods as already demonstrated in the context of sparse Gaussian process (GP) regression [8]. However, due to the nonconvexity of (1) with respect to B, it is susceptible to local minima and accordingly a good heuristic is required to initialize the solution. In this paper, we use a combination of KMP and gradient descent. The basic idea is to assume that at the nth step of KMP, the chosen basis point b n plus the accumulation of basis points obtained until the (n 1)th step (B n 1 ) is a good initial point. Then, at each step of KMP, B n can be subsequently optimized by gradient descent. Naive implementation of this idea is still very expensive. To reduce further the complexity, the following simplifications are adopted: 1. In the KMP step, instead of evaluating the whole training set for choosing b n, only l c (l c l) points are considered; 2. Gradient descent of B n (M) and corresponding A 4 (1:n,:) are performed only at the every rth KMP step. Instead, for each KMP step, only b n and A (n,:) are optimized. In this case, the gradient can be evaluated at O(M l). 5 4 With a slight abuse of the Matlab notation, A (m:n,:) stands for the submatrix of A obtained by extracting the rows of A from m to n. 5 Similarly to [4], A(n) can be analytically calculated at O(M l)cost: A (n) = K bx(n,:)(y K bx(1:n 1,:)A (1:n 1,:) ) λk nb A (1:n 1,:) K bx(n,:) K bx(n,:) + λ. (6)
5 5 At the nth step, the l c candidate basis points for KMP is selected based on a rather cheap criterion: we use the difference between the function output obtained at the (n 1)th step and the estimated desired response of full KRR for each training data points which is then approximated by the localized KRR: for a training data point x i, its NNs are collected in the training set and the full KRR is trained based on only these NNs. The output of this localized KRR for x i gives the estimation of the desired response for x i. It should be noted that these local KRRs cannot be directly applied for regression as they might interpolate poorly on nontraining data points. Once computed at the beginning, the estimated desired responses are fixed throughout the whole optimization process. To gain an insight into the performances of different sparse solution method, a set of preliminary experiments has been performed with KMP, gradient descent (with basis initialized by kmeans algorithm), and the proposed combination of KMP and gradient descent with 10,000 training data points. Figure 1 summarizes the results. Both gradient descent methods outperform KMP, while the combination with KMP provides a better performance. This could be attributed to the better initialization of the solution for the subsequent gradient descent step. 58 KMP Gradient descent KMP+gradient descent cost # basis points Fig. 1. Performance of the different sparse solution methods evaluated in terms of the cost functional (1). A fixed set of hyperparameters were used such that the comparison can be made directly in (1). Combining Candidates. It is possible to construct a superresolved image based on only the scalarvalued regression (i.e., N = 1). However, we propose to predict a patchvalued output such that for each pixel, N different candidates are generated. These candidates constitutes a 3D image Z where the third dimension corresponds the candidates. This setting is motivated by the observation that 1. by sharing the hyperparameters, the computational complexity of resulting patchvalued learning reduces to the scalarvalued learning; 2. the candidates contain information of different input image locations which are actually diverse enough such that the combination can boost the performance: in our preliminary experiments, constructing an image by choosing the best and the worst
6 6 (in terms of the distance to the ground truth) candidates from each 2D location of Z resulted in an average signaltonoise ratio (SNR) difference of 8.24dB. Certainly, the ground truth is not available at actual superresolution stage and accordingly a way of constructing a single pixel out of N candidates is required. One straightforward way is to construct the final estimation as a convex combination of candidates based on a certain confidence measure. For instance, by noting that the (sparse) KRR corresponds to the maximum a posteriori estimation with the (sparse) GP prior [8], one could utilize the predictive variance as a basis for the selection. In the preliminary experiments this resulted in an improvement over the scalarvalued regression. However, a better prediction was obtained when the confidence estimation is obtained based not only on the input patches but also on the context of neighbor reconstructions. For this, a set of linear regressors is trained such that for each location (x, y), they receive a patch of output images Z (NL (x,y),:) and produce the estimation of differences ({d 1 (x, y),..., d N (x, y)}) between the unknown desired output and each candidate. The final estimation of pixel value for an image location (x, y) is then obtained as the convex combination of candidates given in the form of a softmax: Y (x, y) = w i (x, y)z(x, y, i), (7) i=1,...,n where w i (x, y) = exp ( di(x,y) ) [ dj(x,y) σ C / j=1,...,n exp( σ C ) ]. For the experiments in this paper, we set M = 49(7 7), N = 25(5 5), L = 49(7 7), σ k = 0.025, σ C = 0.03, and λ = The values are obtained based on a set of separate validation images. The number of basis points for KRR (l b ) is determined to be 300 as the trade off between the accuracy and the time complexity. In the superresolution experiments, the combination of candidates based on these parameters resulted in an average SNR increase of 0.43dB over the scalarvalued regression. Postprocessing Based on Image Prior. As demonstrated in Fig. 2.b, the result of the proposed regressionbased method is significantly better than the bicubic interpolation. However, detailed visual inspection along the major edges (edges showing rapid and strong change of pixel values) reveals ringing artifacts (oscillation occurred along the edges). In general, regularization methods (depending on the specific class of regularizer) including KRR and SVR tend to fit the data with a smooth function. Accordingly, at the sharp changes of the function (edges in the case of images) oscillation occurs to compensate the resulting loss of smoothness. While this problem can indirectly be resolved by imposing less regularization at the vicinity of edges, more direct approach is to rely on the prior knowledge of discontinuity of images. In this work, we use a modification of the natural image prior (NIP) framework proposed by Tappen et al. [9]: P ({x} {y}) = 1 [ ( ) α ] ˆxi ˆx j exp [ ( ) ] 2 ˆxi y i exp, (8) C (i,j N S (i)) σ N where {y} represents the observed variables corresponding to the pixel values of Y, {x} represents the latent variable, and N S (i) stands for the 8connected i σ R
7 7 neighbors of the pixel location i. While the second product term has the role of preventing the final solution flowing far away from the input regression result Y, the first product term tends to smooth the image based on the costs ˆx i ˆx j. The role of α(< 1) is to reweight the costs such that the largest difference is stressed relatively less than the others such that large changes of pixel values are relatively less penalized. Furthremore, the cost term ˆx i ˆx j α becomes piecewise concave with extreme points at N S (i) such that if the second term is removed, the maximum probability for a pixel i is achieved by assigning it with the value of a neighbor, rather than a certain weighted average of neighbors which might have been the case when α > 1. Accordingly, this distribution prefers a strong edge rather than a set of small edges and can be used to resolve the problem of smoothing around major edges. The optimization of (8) is performed by belief propagation (BP) similarly to [9]. To facilitate the optimization, we reuse the candidate set generated from the regression step such that the best candidates are chosen by BP. a b c d e f Fig. 2. Example of super resolution: a. bicubic, b regression result. c. postprocessed result of b based on NIP, d. Laplacian of bicubic with major edges displayed as green pixels, and e and f. enlarged portions of ac from left to right. Optimizing (8) throughout the whole image region can lead to degraded results as it tends to flatten the textured area, especially, when the contrast is low such that the contribution of the second term is small. 6 This problem is resolved by applying the (modification of) NIP only at the vicinity of major edges. Based on the observation that the input images are blurred and accordingly very high spatial frequency components are removed, the major edges are found by thresholding each pixel of Laplacian of the input image using L 2 and L norms of the local patches encompassing it. It should be noted that the major edge is in general different from the object contour. For instance, in Fig. 2.d, the bound 6 In original work of Tappen et al. [9], this problem does not happen as the candidates are 2 2size image patches rather than individual pixels.
8 8 ary between the chest of the duck and water is not detected as major edges as the intensity variations are not significant across the boundary. In this case, no visible oscillation of pixel values are observed in the original regression result. The parameters α, σ N, and σ R are determined at 0.85, 200 and 1, respectively. While the improvement in terms of SNR is less significant (on average 0.04dB from the combined regression result) the improved visual quality at major edges demonstrate the effectiveness of NIP (Fig. 2). 3 Experiments The proposed method was evaluated based on a set of high and lowresolution image pairs (Fig. 3) which is disjoint from the training images. The desired resolution is twice the input image along each dimension. The number of training data points is 200,000 where it took around a day to train the sparse KRR on a 2.5GHz PC. For comparison, several different examplebased image superresolution methods were evaluated, which include Freeman et al. s NNbased method [2], Tappen et al. s NIP [9], 7 and Kim et al. s SVRbased method [6] (trained based on only 10,000 data points). Fig. 3. Thumbnails of test images: the images are indexed by numbers arranged in the raster order. Figure 4 shows examples of superresolution results. All the examplebased superresolution methods outperform the bicubic interpolation in terms of visual plausibility. The NNbased method and the original NIP produced sharper images at the expense of introducing noise which, even with the improved visual quality, lead to lower SNR values than the bicubic interpolations. The SVR produced less noisy images. However it generated smoothed edges and perceptually distracting ring artifacts which have disappeared for the proposed method. Disregarding the postprocessing stage, we measured on average 0.69dB improvement of SNRs for the proposed method from the SVR. This could be attributed to the sparsity of the solution which enabled training on a large data set and the 7 The original NIP algorithm was developed for superresolving the NNsubsampled image (not bicubic resampling which is used for experiments with all the other methods). Accordingly, for the experiments with NIP, the low resolution images were generated by NN subsampling. The visual qualities of the superresolution results are not significantly different from the results obtained from bicubic resampling. However, the quantitative results should not be directly compared with other methods.
9 9 effectiveness of the candidate combination scheme. Moreover, in comparison to SVR the proposed method requires much less processing time: superresolving a size image into requires around 25 seconds for the proposed method and 20 minutes for the SVRbased method. For quantitative comparison, SNRs of different algorithms are plotted in Fig. 5. a b c d e f g h i j k l Fig. 4. Results of different superresolution algorithms on two images from Fig. 3: ab. original, cd. bicubic, ef. SVR [6], gh. NNbased method [2], ij. NIP [9], and kl. proposed method. 4 Conclusion This paper approached the problem of image superresolution from a nonlinear regression viewpoint. A combination of KMP and gradient descent is adopted to obtain a sparse KRR solution which enabled a realistic application of regressionbased superresolution. To resolve the problem of smoothing artifacts that occur due to the regularization, the NIP was adopted to postprocess the regression result such that the edges are sharpen while the artifacts are suppressed. Comparison with the existing examplebased image superresolution methods demonstrated the effectiveness of the proposed method. Future work should include comparison and combination of various nonexamplebased approaches.
10 10 increase of SNR from bicubic bicubic SVR NN NIP proposed image index Fig. 5. Performance of different superresolutions algorithms. Acknowledgment. The contents of this paper have greatly benefited from discussions with G. BakIr and C. Walder, and comments from anonymous reviewers. The idea of using localized KRR was originated by C. Walder. References 1. Baker, S., Kanade, T.: Limits on superresolution and how to break them. IEEE Trans. Pattern Analysis and Machine Intelligence 24(9), (2002) 2. Freeman, W.T., Jones, T.R., Pasztor, E.C.: Examplebased superresolution. IEEE Computer Graphics and Applications 22(2), (2002) 3. Hertzmann, A., Jacobs, C.E., Oliver, N., Curless, B., Salesin, D.H.: Image analogies. In: Computer Graphics (Proc. Siggraph 2001), pp ACM Press, NY (2001) 4. Keerthi, S.S., Chu, W.: A matching pursuit approach to sparse gaussian process regression. In: Advances in Neural Information Processing Systems. MIT Press, Cambridge, MA (2005) 5. Kim, K.I., Franz, M.O., Schölkopf, B.: Iterative kernel principal component analysis for image modeling. IEEE Trans. Pattern Analysis and Machine Intelligence 27(9), (2005) 6. Kim, K.I., Kim, D.H., Kim, J.H.: Examplebased learning for image superresolution. In: Proc. the third TsinghuaKAIST Joint Workshop on Pattern Recognition, pp (2004) 7. Ni, K., Nguyen, T.Q.: Image superresolution using support vector regression. IEEE Trans. Image Processing 16(6), (2007) 8. Snelson, E., Ghahramani, Z.: Sparse gaussian processes using pseudoinputs. In: Advances in Neural Information Processing Systems. MIT Press, Cambridge, MA (2006) 9. Tappen, M.F., Russel, B.C., Freeman, W.T.: Exploiting the sparse derivative prior for superresolution and image demosaicing. In: Proc. IEEE Workshop on Statistical and Computational Theories of Vision (2003) 10. Tschumperlé, D., Deriche, R.: Vectorvalued image regularization with pdes: a common framework for different applications. IEEE Trans. Pattern Analysis and Machine Intelligence 27(4), (2005) 11. Vincent, P., Bengio, Y.: Kernel matching pursuit. Machine Learning 48, (2002)
Distinctive Image Features from ScaleInvariant Keypoints
Distinctive Image Features from ScaleInvariant Keypoints David G. Lowe Computer Science Department University of British Columbia Vancouver, B.C., Canada lowe@cs.ubc.ca January 5, 2004 Abstract This paper
More informationFrom Few to Many: Illumination Cone Models for Face Recognition Under Variable Lighting and Pose. Abstract
To Appear in the IEEE Trans. on Pattern Analysis and Machine Intelligence From Few to Many: Illumination Cone Models for Face Recognition Under Variable Lighting and Pose Athinodoros S. Georghiades Peter
More informationTHE PROBLEM OF finding localized energy solutions
600 IEEE TRANSACTIONS ON SIGNAL PROCESSING, VOL. 45, NO. 3, MARCH 1997 Sparse Signal Reconstruction from Limited Data Using FOCUSS: A Reweighted Minimum Norm Algorithm Irina F. Gorodnitsky, Member, IEEE,
More informationSuperResolution Through Neighbor Embedding
SuperResolution Through Neighbor Embedding Hong Chang, DitYan Yeung, Yimin Xiong Department of Computer Science Hong Kong University of Science and Technology Clear Water Bay, Kowloon, Hong Kong {hongch,
More informationObject Detection with Discriminatively Trained Part Based Models
1 Object Detection with Discriminatively Trained Part Based Models Pedro F. Felzenszwalb, Ross B. Girshick, David McAllester and Deva Ramanan Abstract We describe an object detection system based on mixtures
More informationHighDimensional Image Warping
Chapter 4 HighDimensional Image Warping John Ashburner & Karl J. Friston The Wellcome Dept. of Imaging Neuroscience, 12 Queen Square, London WC1N 3BG, UK. Contents 4.1 Introduction.................................
More informationObject Segmentation by Long Term Analysis of Point Trajectories
Object Segmentation by Long Term Analysis of Point Trajectories Thomas Brox 1,2 and Jitendra Malik 1 1 University of California at Berkeley 2 AlbertLudwigsUniversity of Freiburg, Germany {brox,malik}@eecs.berkeley.edu
More informationLearning to Select Features using their Properties
Journal of Machine Learning Research 9 (2008) 23492376 Submitted 8/06; Revised 1/08; Published 10/08 Learning to Select Features using their Properties Eyal Krupka Amir Navot Naftali Tishby School of
More informationTHE adoption of classical statistical modeling techniques
236 IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, VOL. 28, NO. 2, FEBRUARY 2006 Data Driven Image Models through Continuous Joint Alignment Erik G. LearnedMiller Abstract This paper
More informationGenerative or Discriminative? Getting the Best of Both Worlds
BAYESIAN STATISTICS 8, pp. 3 24. J. M. Bernardo, M. J. Bayarri, J. O. Berger, A. P. Dawid, D. Heckerman, A. F. M. Smith and M. West (Eds.) c Oxford University Press, 2007 Generative or Discriminative?
More informationLearning Invariant Features through Topographic Filter Maps
Learning Invariant Features through Topographic Filter Maps Koray Kavukcuoglu Marc Aurelio Ranzato Rob Fergus Yann LeCun Courant Institute of Mathematical Sciences New York University {koray,ranzato,fergus,yann}@cs.nyu.edu
More informationGenerating more realistic images using gated MRF s
Generating more realistic images using gated MRF s Marc Aurelio Ranzato Volodymyr Mnih Geoffrey E. Hinton Department of Computer Science University of Toronto {ranzato,vmnih,hinton}@cs.toronto.edu Abstract
More informationDiscovering objects and their location in images
Discovering objects and their location in images Josef Sivic Bryan C. Russell Alexei A. Efros Andrew Zisserman William T. Freeman Dept. of Engineering Science CS and AI Laboratory School of Computer Science
More informationFeature Sensitive Surface Extraction from Volume Data
Feature Sensitive Surface Extraction from Volume Data Leif P. Kobbelt Mario Botsch Ulrich Schwanecke HansPeter Seidel Computer Graphics Group, RWTHAachen Computer Graphics Group, MPI Saarbrücken Figure
More informationSubspace Pursuit for Compressive Sensing: Closing the Gap Between Performance and Complexity
Subspace Pursuit for Compressive Sensing: Closing the Gap Between Performance and Complexity Wei Dai and Olgica Milenkovic Department of Electrical and Computer Engineering University of Illinois at UrbanaChampaign
More informationLearning Deep Architectures for AI. Contents
Foundations and Trends R in Machine Learning Vol. 2, No. 1 (2009) 1 127 c 2009 Y. Bengio DOI: 10.1561/2200000006 Learning Deep Architectures for AI By Yoshua Bengio Contents 1 Introduction 2 1.1 How do
More informationEfficient Learning of Sparse Representations with an EnergyBased Model
Efficient Learning of Sparse Representations with an EnergyBased Model Marc Aurelio Ranzato Christopher Poultney Sumit Chopra Yann LeCun Courant Institute of Mathematical Sciences New York University,
More informationRegression. Chapter 2. 2.1 Weightspace View
Chapter Regression Supervised learning can be divided into regression and classification problems. Whereas the outputs for classification are discrete class labels, regression is concerned with the prediction
More informationAn Experimental Comparison of MinCut/MaxFlow Algorithms for Energy Minimization in Vision
In IEEE Transactions on PAMI, Vol. 26, No. 9, pp. 11241137, Sept. 2004 p.1 An Experimental Comparison of MinCut/MaxFlow Algorithms for Energy Minimization in Vision Yuri Boykov and Vladimir Kolmogorov
More informationRobust Object Detection with Interleaved Categorization and Segmentation
Submission to the IJCV Special Issue on Learning for Vision and Vision for Learning, Sept. 2005, 2 nd revised version Aug. 2007. Robust Object Detection with Interleaved Categorization and Segmentation
More informationSumProduct Networks: A New Deep Architecture
SumProduct Networks: A New Deep Architecture Hoifung Poon and Pedro Domingos Computer Science & Engineering University of Washington Seattle, WA 98195, USA {hoifung,pedrod}@cs.washington.edu Abstract
More informationScalable Collaborative Filtering with Jointly Derived Neighborhood Interpolation Weights
Seventh IEEE International Conference on Data Mining Scalable Collaborative Filtering with Jointly Derived Neighborhood Interpolation Weights Robert M. Bell and Yehuda Koren AT&T Labs Research 180 Park
More informationFlexible and efficient Gaussian process models for machine learning
Flexible and efficient Gaussian process models for machine learning Edward Lloyd Snelson M.A., M.Sci., Physics, University of Cambridge, UK (2001) Gatsby Computational Neuroscience Unit University College
More informationRandomized Trees for RealTime Keypoint Recognition
Randomized Trees for RealTime Keypoint Recognition Vincent Lepetit Pascal Lagger Pascal Fua Computer Vision Laboratory École Polytechnique Fédérale de Lausanne (EPFL) 1015 Lausanne, Switzerland Email:
More informationGaussian Process Dynamical Models
Gaussian Process Dynamical Models Jack M. Wang, David J. Fleet, Aaron Hertzmann Department of Computer Science University of Toronto, Toronto, ON M5S 3G4 {jmwang,hertzman}@dgp.toronto.edu, fleet@cs.toronto.edu
More informationFace Alignment by Explicit Shape Regression
Face Alignment by Explicit Shape Regression Xudong Cao Yichen Wei Fang Wen Jian Sun Microsoft Research Asia {xudongca,yichenw,fangwen,jiansun}@microsoft.com Abstract. We present a very efficient, highly
More informationThe Pyramid Match Kernel: Discriminative Classification with Sets of Image Features
The Pyramid Match Kernel: Discriminative Classification with Sets of Image Features Kristen Grauman and Trevor Darrell Massachusetts Institute of Technology Computer Science and Artificial Intelligence
More informationAn Introduction to Variable and Feature Selection
Journal of Machine Learning Research 3 (23) 11571182 Submitted 11/2; Published 3/3 An Introduction to Variable and Feature Selection Isabelle Guyon Clopinet 955 Creston Road Berkeley, CA 9478151, USA
More informationUnsupervised Spike Detection and Sorting with Wavelets and Superparamagnetic Clustering
LETTER Communicated by Maneesh Sahani Unsupervised Spike Detection and Sorting with Wavelets and Superparamagnetic Clustering R. Quian Quiroga rodri@vis.caltech.edu Z. Nadasdy zoltan@vis.caltech.edu Division
More informationA Unifying View of Sparse Approximate Gaussian Process Regression
Journal of Machine Learning Research 6 (2005) 1939 1959 Submitted 10/05; Published 12/05 A Unifying View of Sparse Approximate Gaussian Process Regression Joaquin QuiñoneroCandela Carl Edward Rasmussen
More information