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

A compressive sensing-based multispectral video imager comprises a beamsplitter, a first light channel, a second light channel, and an image reconstruction processor; the beamsplitter is configured to divide the beam of an underlying image into a first light beam and a second light beam; the first light beam enters the first light channel, processed and sampled in the first light channel, to obtain a first mixing spectral image which is transferred to the image reconstruction processor; the second light beam enters the second light channel, processed and sampled in the second light channel, to obtain a second mixing spectral image which is transferred to the image reconstruction processor; the image reconstruction processor is configured to reconstruct the underlying spectral image based on the first mixing spectral image and the second mixing spectral image by using non-linear optimization method.

Dual-channel multispectral video imaging device and imaging method based on compressive sensing

The increasing illegal parking has become more and more serious. Nowadays the methods of detecting illegally parked vehicles are based on background segmentation. However, this method is weakly robust and sensitive to environment. Benefitting from deep learning, this paper proposes a novel illegal vehicle parking detection system. Illegal vehicles captured by camera are firstly located and classified by the famous Single Shot MultiBox Detector (SSD) algorithm. To improve the performance, we propose to optimize SSD by adjusting the aspect ratio of default box to accommodate with our dataset better. After that, a tracking and analysis of movement is adopted to judge the illegal vehicles in the region of interest (ROI). Experiments show that the system can achieve a 99% accuracy and real-time (25FPS) detection with strong robustness in complex environments.

Real-Time Illegal Parking Detection System Based on Deep Learning

The majority of the existing work on designing nonuniform filter banks (NUFBs) cannot achieve linear-phase (LP) property, because of the high complexity associated with phase distribution. This correspondence proposes an idea of partial cosine modulation to obtain the LP property of NUFBs with rational sampling factors. It makes the efficient modulation technique possible to be used to design LP NUFBs. Except the separately designed lowpass and highpass analysis/synthesis filters, we obtain the bandpass by cosine modulation of several prototypes, worth of the name ?partial cosine modulation.? By analyzing the phase issue of significant aliasing terms, we derive the matching conditions to make the required LP NUFB achievable. With these criteria being satisfied, the design problem becomes that of several prototypes as well as the lowpass and highpass filters, leading to a less design effort. By using the proposed method, near-perfect-reconstruction LP NUFBs can be obtained in a simple and efficient way as demonstrated by examples.

/pdf/design-of-linear-phase-nonuniform-filter-banks-with-partial-35s8w6rs27.pdf

Design of Linear-Phase Nonuniform Filter Banks With Partial Cosine Modulation

Network pruning has been proposed as a remedy for alleviating the over-parameterization problem of deep neural networks. However, its value has been recently challenged especially from the perspective of neural architecture search (NAS). We challenge the conventional wisdom of pruningafter-training by proposing a joint search-andtraining approach that directly learns a compact network from the scratch. By treating pruning as a search strategy, we present two new insights in this paper: 1) it is possible to expand the search space of networking pruning by associating each filter with a learnable weight; 2) joint search-and-training can be conducted iteratively to maximize the learning efficiency. More specifically, we propose a coarseto-fine tuning strategy to iteratively sample and update compact sub-network to approximate the target network. The weights associated with network filters will be accordingly updated by joint searchand-training to reflect learned knowledge in NAS space. Moreover, we introduce strategies of random perturbation (inspired by Monte Carlo) and flexible thresholding (inspired by Reinforcement Learning) to adjust the weight and size of each layer. Extensive experiments on ResNet and VGGNet demonstrate the superior performance of our proposed method on popular datasets including CIFAR10, CIFAR100 and ImageNet.

/pdf/beyond-network-pruning-a-joint-search-and-training-approach-4ubxhozsvh.pdf

Beyond Network Pruning: a Joint Search-and-Training Approach

Physical acquisition of high-resolution hyperspectral images (HR-HSI) has remained difficult, despite its potential of resolving material-related ambiguities in vision applications. Deep hyperspectral image fusion, aiming at reconstructing an HR-HSI from a pair of low-resolution hyperspectral image (LR-HSI) and high-resolution multispectral image (HR-MSI), has become an appealing computational alternative. Existing fusion methods either rely on hand-crafted image priors or treat fusion as a nonlinear mapping problem, ignoring important physical imaging models. In this paper, we propose a novel regularization strategy to fully exploit the spatio-spectral dependency by a spatially adaptive 3D filter. Moreover, the joint exploitation of spatio-spectral regularization and physical imaging models inspires us to formulate deep hyperspectral image fusion as a differentiable optimization problem. We show how to solve this optimization problem by an end-to-end training of a model-guided unfolding network named DHIF-Net. Unlike existing works of simply concatenating spatial with spectral regularization, our approach aims at an end-to-end optimization of iterative spatio-spectral regularization by multistage network implementations. Our extensive experimental results on both synthetic and real datasets have shown that our DHIF-Net outperforms other competing methods in terms of both objective and subjective visual quality.

Guangming Shi

Papers

Dual-channel multispectral video imaging device and imaging method based on compressive sensing

Real-Time Illegal Parking Detection System Based on Deep Learning

Design of Linear-Phase Nonuniform Filter Banks With Partial Cosine Modulation

Beyond Network Pruning: a Joint Search-and-Training Approach

Deep Hyperspectral Image Fusion Network With Iterative Spatio-Spectral Regularization