We propose a deep convolutional neural network architecture codenamed Inception that achieves the new state of the art for classification and detection in the ImageNet Large-Scale Visual Recognition Challenge 2014 (ILSVRC14). The main hallmark of this architecture is the improved utilization of the computing resources inside the network. By a carefully crafted design, we increased the depth and width of the network while keeping the computational budget constant. To optimize quality, the architectural decisions were based on the Hebbian principle and the intuition of multi-scale processing. One particular incarnation used in our submission for ILSVRC14 is called GoogLeNet, a 22 layers deep network, the quality of which is assessed in the context of classification and detection.

/pdf/going-deeper-with-convolutions-1yobw2o2ds.pdf

Going deeper with convolutions

[신간의 별자리x] 우리/미술, 그리고 ‘슬픔의 박물관’

Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

I have developed "tennis elbow" from lugging this book around the past four weeks, but it is worth the pain, the effort, and the aspirin. It is also worth the (relatively speaking) bargain price. Including appendixes, this book contains 894 pages of text. The entire panorama of the neural sciences is surveyed and examined, and it is comprehensive in its scope, from genomes to social behaviors. The editors explicitly state that the book is designed as "an introductory text for students of biology, behavior, and medicine," but it is hard to imagine any audience, interested in any fragment of neuroscience at any level of sophistication, that would not enjoy this book. The editors have done a masterful job of weaving together the biologic, the behavioral, and the clinical sciences into a single tapestry in which everyone from the molecular biologist to the practicing psychiatrist can find and appreciate his or

Principles of Neural Science

Despite the breakthroughs in accuracy and speed of single image super-resolution using faster and deeper convolutional neural networks, one central problem remains largely unsolved: how do we recover the finer texture details when we super-resolve at large upscaling factors? The behavior of optimization-based super-resolution methods is principally driven by the choice of the objective function. Recent work has largely focused on minimizing the mean squared reconstruction error. The resulting estimates have high peak signal-to-noise ratios, but they are often lacking high-frequency details and are perceptually unsatisfying in the sense that they fail to match the fidelity expected at the higher resolution. In this paper, we present SRGAN, a generative adversarial network (GAN) for image super-resolution (SR). To our knowledge, it is the first framework capable of inferring photo-realistic natural images for 4x upscaling factors. To achieve this, we propose a perceptual loss function which consists of an adversarial loss and a content loss. The adversarial loss pushes our solution to the natural image manifold using a discriminator network that is trained to differentiate between the super-resolved images and original photo-realistic images. In addition, we use a content loss motivated by perceptual similarity instead of similarity in pixel space. Our deep residual network is able to recover photo-realistic textures from heavily downsampled images on public benchmarks. An extensive mean-opinion-score (MOS) test shows hugely significant gains in perceptual quality using SRGAN. The MOS scores obtained with SRGAN are closer to those of the original high-resolution images than to those obtained with any state-of-the-art method.

/pdf/photo-realistic-single-image-super-resolution-using-a-2y0438rlmd.pdf

Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network

Conventional wisdom in blind super-resolution (SR) first estimates the unknown degradation from the low-resolution image and then exploits the degradation information for image reconstruction. Such sequential approaches suffer from two fundamental weaknesses - i.e., the lack of robustness (the performance drops when the estimated degradation is inaccurate) and the lack of transparency (network architectures are heuristic without incorporating domain knowledge). To address these issues, we propose a joint Maximum a Posteriori (MAP) approach for estimating the unknown kernel and high-resolution image simultaneously. Our method first introduces uncertainty learning in the latent space when estimating the blur kernel, aiming at improving the robustness to the estimation error. Then we propose a novel SR network by unfolding the joint MAP estimator with a learned Laplacian Scale Mixture (LSM) prior and the estimated kernel. We have also developed a novel approach of estimating both the scale prior coefficient and the local means of the LSM model through a deep convolutional neural network (DCNN). All parameters of the MAP estimation algorithm and the DCNN parameters are jointly optimized through end-to-end training. Extensive experiments on both synthetic and real-world images show that our method achieves state-of-the-art performance for the task of blind image SR. 

Uncertainty Learning in Kernel Estimation for Multi-stage Blind Image Super-Resolution

Image registration is a fundamental image processing task to match and align physically two images which could have been imaged by different sensors, view angles and or at different times. This paper presents a new algorithm for image registration based on a combination between cross correlation function and feature-based methods .In the proposed algorithm, the ground control points are extracted from input images using the Harris corner detector, and the correspondence between the points extracted from the different images is established using RANdom SAmple Consensus (RANSAC) method with. The image coregistration in this paper is performed by three ways: It is first performed by the proposed algorithm then it is performed by feature-based methods and finally the results are compared with those achieved using Gamma design software. The method is also evaluated by using the coherence map. The results showed that the combination between cross correlation function and feature-based methods almost give us the same results of traditional method but with low processing time., which is our aim to prove and validate. (4 pages)

Semi-automatic image registration using harris corner detection and RANdom SAmple Consensus (RANSAC)

As a fundamental role in radar signal processing, direction of arrival (DOA) estimation has drawn much attention from researchers. This paper introduces a new wideband DOA estimation model. In this model, instead of processing each subband individually, the whole frequency information of the received signal is combined. And, to utilize the same sparsity among angle spectrums with different subbands, a novel sparsity constraint is proposed. Based on sparse representation framework, an optimization is constructed and solved by second-order cone (SOC) programming to obtain the DOA information. Simulations are provided to prove the superior performance of the proposed method.

Wideband DOA estimation based on sparse representation — An extension of l 1 -SVD in wideband cases

As an emerging field, Automated Machine Learning (AutoML) aims to reduce or eliminate manual operations that require expertise in machine learning. In this paper, a graph-based architecture is employed to represent flexible combinations of ML models, which provides a large searching space compared to tree-based and stacking-based architectures. Based on this, an evolutionary algorithm is proposed to search for the best architecture, where the mutation and heredity operators are the key for architecture evolution. With Bayesian hyper-parameter optimization, the proposed approach can automate the workflow of machine learning. On the PMLB dataset, the proposed approach shows the state-of-the-art performance compared with TPOT, Autostacker, and auto-sklearn. Some of the optimized models are with complex structures which are difficult to obtain in manual design.

DarwinML: A Graph-based Evolutionary Algorithm for Automated Machine Learning.

The orthogonal signals of multi-carrier-frequency emission and multiple antennas receipt module are used in SIAR radar. The corresponding received echo is equivalent to non-uniform spatial sampling after the frequency diversity process. As using the traditional Fourier transform will result in the target spectral with large sidelobe, the method presented in this paper firstly makes the preordering treatment for the position of the received antenna. Then, the Bayesian maximum posteriori estimation with l2-norm weighted constraint is utilized to achieve the equivalent uniform array echo. The simulations present the spectrum estimation in angle precision estimation of multiple targets under different SNRs, different virtual antenna numbers and different elevations. The estimation results confirm the advantage of SIAR radar both in array expansion and angle estimation.

Guangming Shi

Papers

Uncertainty Learning in Kernel Estimation for Multi-stage Blind Image Super-Resolution

Semi-automatic image registration using harris corner detection and RANdom SAmple Consensus (RANSAC)

Wideband DOA estimation based on sparse representation — An extension of l 1 -SVD in wideband cases

DarwinML: A Graph-based Evolutionary Algorithm for Automated Machine Learning.

Angel estimation via frequency diversity of the SIAR radar based on Bayesian theory