Neural Networks In Computer Vision Research Articles

Novel Convergence Results of Adaptive Stochastic Gradient Descents.

Adaptive stochastic gradient descent, which uses unbiased samples of the gradient with stepsizes chosen from the historical information, has been widely used to train neural networks for computer vision and pattern recognition tasks. This paper revisits the theoretical aspects of two classes of adaptive stochastic gradient descent methods, which contain several existing state-of-the-art schemes. We focus on the presentation of novel findings: In the general smooth case, the nonergodic convergence results are given, that is, the expectation of the gradients' norm rather than the minimum of past iterates is proved to converge; We also studied their performances under Polyak-Łojasiewicz property on the objective function. In this case, the nonergodic convergence rates are given for the expectation of the function values. Our findings show that more substantial restrictions on the steps are needed to guarantee the nonergodic function values' convergence (rates).

Comparative Study of End-to-end Deep Learning Methods for Self-driving Car

Self-driving car is one of the most amazing applications and most active research of artificial intelligence. It uses end-to-end deep learning models to take orientation and speed decisions, using mainly Convolutional Neural Networks for computer vision, plugged to a fully connected network to output control commands. In this paper, we introduce the Self-driving car domain and the CARLA simulation environment with a focus on the lane-keeping task, then we present the two main end-to-end models, used to solve this problematic, beginning by Deep imitation learning (IL) and specifically the Conditional Imitation Learning (COIL) algorithm, that learns through expert labeled demonstrations, trying to mimic their behaviors, and thereafter, describing Deep Reinforcement Learning (DRL), and precisely DQN and DDPG (respectively Deep Q learning and deep deterministic policy gradient), that uses the concepts of learning by trial and error, while adopting the Markovian decision processes (MDP), to get the best policy for the driver agent. In the last chapter, we compare the two algorithms IL and DRL based on a new approach, with metrics used in deep learning (Loss during training phase) and Self-driving car (the episode's duration before a crash and Average distance from the road center during the testing phase). The results of the training and testing on CARLA simulator reveals that the IL algorithm performs better than DRL algorithm when the agents are already trained on a given circuit, but DRL agents show better adaptability when they are on new roads.

Encoder-Decoder Full Residual Deep Networks for Robust Regression and Spatiotemporal Estimation.

Although increasing hidden layers can improve the ability of a neural network in modeling complex nonlinear relationships, deep layers may result in degradation of accuracy due to the problem of vanishing gradient. Accuracy degradation limits the applications of deep neural networks to predict continuous variables with a small sample size and/or weak or little invariance to translations. Inspired by residual convolutional neural network in computer vision, we developed an encoder-decoder full residual deep network to robustly regress and predict complex spatiotemporal variables. We embedded full shortcuts from each encoding layer to its corresponding decoding layer in a systematic encoder-decoder architecture for efficient residual mapping and error signal propagation. We demonstrated, theoretically and experimentally, that the proposed network structure with full residual connections can successfully boost the backpropagation of signals and improve learning outcomes. This novel method has been extensively evaluated and compared with four commonly used methods (i.e., plain neural network, cascaded residual autoencoder, generalized additive model, and XGBoost) across different testing cases for continuous variable predictions. For model evaluation, we focused on spatiotemporal imputation of satellite aerosol optical depth with massive nonrandomness missingness and spatiotemporal estimation of atmospheric fine particulate matter [Formula: see text] (PM2.5). Compared with the other approaches, our method achieved the state-of-the-art accuracy, had less bias in predicting extreme values, and generated more realistic spatial surfaces. This encoder-decoder full residual deep network can be an efficient and powerful tool in a variety of applications that involve complex nonlinear relationships of continuous variables, varying sample sizes, and spatiotemporal data with weak or little invariance to translation.

Application Research of Deep Convolutional Neural Network in Computer Vision

<p>As an important research achievement in the field of brain like computing, deep convolution neural network has been widely used in many fields such as computer vision, natural language processing, information retrieval, speech recognition, semantic understanding and so on. It has set off a wave of neural network research in industry and academia and promoted the development of artificial intelligence. At present, the deep convolution neural network mainly simulates the complex hierarchical cognitive laws of the human brain by increasing the number of layers of the network, using a larger training data set, and improving the network structure or training learning algorithm of the existing neural network, so as to narrow the gap with the visual system of the human brain and enable the machine to acquire the capability of "abstract concepts". Deep convolution neural network has achieved great success in many computer vision tasks such as image classification, target detection, face recognition, pedestrian recognition, etc. Firstly, this paper reviews the development history of convolutional neural networks. Then, the working principle of the deep convolution neural network is analyzed in detail. Then, this paper mainly introduces the representative achievements of convolution neural network from the following two aspects, and shows the improvement effect of various technical methods on image classification accuracy through examples. From the aspect of adding network layers, the structures of classical convolutional neural networks such as AlexNet, ZF-Net, VGG, GoogLeNet and ResNet are discussed and analyzed. From the aspect of increasing the size of data set, the difficulties of manually adding labeled samples and the effect of using data amplification technology on improving the performance of neural network are introduced. This paper focuses on the latest research progress of convolution neural network in image classification and face recognition. Finally, the problems and challenges to be solved in future brain-like intelligence research based on deep convolution neural network are proposed.</p>

An Interactive Visualization for Feature Localization in Deep Neural Networks.

Deep artificial neural networks have become the go-to method for many machine learning tasks. In the field of computer vision, deep convolutional neural networks achieve state-of-the-art performance for tasks such as classification, object detection, or instance segmentation. As deep neural networks become more and more complex, their inner workings become more and more opaque, rendering them a “black box” whose decision making process is no longer comprehensible. In recent years, various methods have been presented that attempt to peek inside the black box and to visualize the inner workings of deep neural networks, with a focus on deep convolutional neural networks for computer vision. These methods can serve as a toolbox to facilitate the design and inspection of neural networks for computer vision and the interpretation of the decision making process of the network. Here, we present the new tool Interactive Feature Localization in Deep neural networks (IFeaLiD) which provides a novel visualization approach to convolutional neural network layers. The tool interprets neural network layers as multivariate feature maps and visualizes the similarity between the feature vectors of individual pixels of an input image in a heat map display. The similarity display can reveal how the input image is perceived by different layers of the network and how the perception of one particular image region compares to the perception of the remaining image. IFeaLiD runs interactively in a web browser and can process even high resolution feature maps in real time by using GPU acceleration with WebGL 2. We present examples from four computer vision datasets with feature maps from different layers of a pre-trained ResNet101. IFeaLiD is open source and available online at https://ifealid.cebitec.uni-bielefeld.de.

Adversarial attack on DL-based massive MIMO CSI feedback

With the increasing application of deep learning (DL) algorithms in wireless communications, the physical layer faces new challenges caused by adversarial attack. Such attack has significantly affected the neural network in computer vision. We choose DL-based channel state information (CSI) to show the effect of adversarial attack on DL-based communication system. We present a practical method to craft white-box adversarial attack on DL-based CSI feedback process. Our simulation results show the destructive effect adversarial attack causes on DL-based CSI feedback by analyzing the performance of normalized mean square error. We also launch a jamming attack for comparison and find that the jamming attack could be prevented with certain precautions. As DL algorithm becomes the trend in developing wireless communication, this work raises concerns regarding the security in the use of DL-based algorithms.

AGE ESTIMATION USING SPECIFIC DOMAIN TRANSFER LEARNING

Nowadays, the engagement of deep neural networks in computer vision increases the ability to achieve higher accuracy in many learning tasks such as face recognition and detection. However, the automatic estimation of human age is still considered as the most challenging facial task that demands extra efforts to obtain an accepted accuracy for real application. In this paper, we attempt to obtain a satisfied model that overcomes the overfitting problem, by fine-tuning CNNs model which was pretrained on face recognition task to estimate the real age. To make the model more robust, we evaluated the model for real age estimation on two types of datasets: on the constrained FG_NET dataset we achieved 3.446 of MAE, on the unconstrained UTKFace dataset we achieved 4.867 MAE. The experimental results of our approach outperform other state-of-the-art age estimation models on the benchmark datasets. We also fine-tuned the model for age groups classification task on Adience dataset, and our model achieved an accuracy of 61.4.

A Novel Semantics-Preserving Hashing for Fine-Grained Image Retrieval

With the advent of the era of big data, the storage and retrieval of data have become a research hotspot. Hashing methods that transform high-dimensional data into compact binary codes have received increasing attention. Recently, with the successful application of convolutional neural networks in computer vision, deep hashing methods utilize an end-to-end framework to learn feature representations and hash codes mutually, which achieve better retrieval performance than conventional hashing methods. However, deep hashing methods still face some challenges in image retrieval. Firstly, most existing deep hashing methods preserve similarity between original data space and hash coding space using loss functions with high time complexity, which cannot get a win-win situation in time and accuracy. Secondly, few existing deep hashing methods are designed for fine-grained image retrieval, which is necessary in practice. In this study, we propose a novel semantics-preserving hashing method which solves the above problems. We add a hash layer before the classification layer as a feature switch layer to guide the classification. At the same time, we replace the complicated loss with the simple classification loss, combining with quantization loss and bit balance loss to generate high-quality hash codes. Besides, we incorporate feature extractor designed for fine-grained image classification into our network for better representation learning. The results on three widely-used fine-grained image datasets show that our method is superior to other state-of-the-art image retrieval methods.

Bibliometric Analysis of the Application of Convolutional Neural Network in Computer Vision

This article analyzes the research progress in field of Convolutional Neural Networks (CNNs) using the bibliometric method. Literature samples of CNNs are analyzed by a basic statistic and co-citation network. Experimental results show that CNNs are being utilized in many computer vision applications, such as fault and image recognition diagnosis, seismic detection, positioning, and automatic detection of cracks and signals, image classification and image segmentation. In addition, there is systematic research on unbalanced problems in CNNs. Quantitative experimental research, extensive application fields, and market research informatization will be the three vital research tendencies in the future. The ideas and conclusions of this article provide insights to the academic research of CNNs and their practical application in the corporate world.

Physics Inspired Deep Neural Networks for Top Quark Reconstruction

Deep neural networks (DNNs) have been applied to the fields of computer vision and natural language processing with great success in recent years. The success of these applications has hinged on the development of specialized DNN architectures that take advantage of specific characteristics of the problem to be solved, namely convolutional neural networks for computer vision and recurrent neural networks for natural language processing. This research explores whether a neural network architecture specific to the task of identifying t → Wb decays in particle collision data yields better performance than a generic, fully-connected DNN. Although applied here to resolved top quark decays, this approach is inspired by an DNN technique for tagging boosted top quarks, which consists of defining custom neural network layers known as the combination and Lorentz layers. These layers encode knowledge of relativistic kinematics applied to combinations of particles, and the output of these specialized layers can then be fed into a fully connected neural network to learn tasks such as classification. This research compares the performance of these physics inspired networks to that of a generic, fully-connected DNN, to see if there is any advantage in terms of classification performance, size of the network, or ease of training.

Improved Traffic Sign Recognition Using Deep ConvNet Architecture

Everyday more and more accidents between vehicles accrue mainly because of poorly visible road signs, that is why it is important to have a system that can help drivers for safer driving. In order to improve road safety and significantly reduce the risk of these accidents, numerous works have been undertaken in computer vision. In this paper, we propose a contribution to this issue by taking advantage of the remarkable results of deep convolutional neural networks in computer vision. We propose an automatic recognition system of road signs based on a modified model inspired by LeNet model. The results obtained by comparison of LeNet model and two proposed modified models on the German traffic dataset is about 99% accuracy which is promising compared to the state-of-the-art results.

Deep attention based music genre classification

As an important component of music information retrieval, music genre classification attracts great attentions these years. Benefitting from the outstanding performance of deep neural networks in computer vision, some researchers apply CNN on music genre classification tasks with audio spectrograms as input instead, which has similarities with RGB images. These methods are based on a latent assumption that spectrums with different temporal steps have equal importance. However, it goes against the theory of processing bottleneck in psychology as well as our observation from audio spectrograms. By considering the differences of spectrums, we propose a new model incorporating with attention mechanism based on Bidirectional Recurrent Neural Network. Furthermore, two attention-based models (serial attention and parallelized attention) are implemented in this paper. Comparing with serial attention, parallelized attention is more flexible and gets better results in our experiments. Especially, the CNN-based parallelized attention models with taking STFT spectrograms as input outperform the previous work.

TableSense: Spreadsheet Table Detection with Convolutional Neural Networks

Spreadsheet table detection is the task of detecting all tables on a given sheet and locating their respective ranges. Automatic table detection is a key enabling technique and an initial step in spreadsheet data intelligence. However, the detection task is challenged by the diversity of table structures and table layouts on the spreadsheet. Considering the analogy between a cell matrix as spreadsheet and a pixel matrix as image, and encouraged by the successful application of Convolutional Neural Networks (CNN) in computer vision, we have developed TableSense, a novel end-to-end framework for spreadsheet table detection. First, we devise an effective cell featurization scheme to better leverage the rich information in each cell; second, we develop an enhanced convolutional neural network model for table detection to meet the domain-specific requirement on precise table boundary detection; third, we propose an effective uncertainty metric to guide an active learning based smart sampling algorithm, which enables the efficient build-up of a training dataset with 22,176 tables on 10,220 sheets with broad coverage of diverse table structures and layouts. Our evaluation shows that TableSense is highly effective with 91.3% recall and 86.5% precision in EoB-2 metric, a significant improvement over both the current detection algorithm that are used in commodity spreadsheet tools and state-of-the-art convolutional neural networks in computer vision.

Vehicle Detection in Aerial Images Based on Lightweight Deep Convolutional Network and Generative Adversarial Network

Vehicle detection in aerial images is a challenging task and plays an important role in a wide range of applications. Traditional detection algorithms are based on sliding-window searching and shallowlearning-based features, which limits the ability to represent features and generates a lot of computational costs. Recently, with the successful application of convolutional neural network in computer vision, many state-of-the-art detectors have been developed based on deep CNNs. However, these CNN-based models still face some difficulties and challenges in vehicle detection in aerial images. Firstly, the CNN-based detection model requires extensive calculations during training and detection, and the accuracy of detection for small objects is not high. In addition, deep learning models often require a large amount of sample data to train a robust detection model, while the annotated data of aerial vehicles is limited. In this study, we propose a lightweight deep convolutional neural network detection model named LD-CNNs. The detection algorithm not only greatly reduces the computational costs of the model, but also significantly improves the accuracy of the detection. What’s more, in order to cope with the problem of insufficient training samples, we develop a multi-condition constrained generative adversarial network named MC-GAN, which can effectively generate samples. The detection performance of the proposed model has been evaluated on the Munich public dataset and the collected dataset respectively. The results show that on the Munich dataset, the proposed method achieves 86.9% on mAP (mean average precision), F1-score is 0.875, and the detection time is 1.64s on Nvidia Titan XP. At present, these detection indicators have reached state-of-the-art level in vehicle detection of aerial images.

Generative Adversarial Network Based Image Augmentation for Insect Pest Classification Enhancement

The application of deep neural networks in computer vision has remarkably improved the reliability of automated insect pest identification algorithms. However, to build a robust deep neural network classifier model, sufficient number of image data is necessary. In preparing image data, collection and annotation of large datasets can be labor-intensive and time-consuming. This work aims to reduce the effort and time needed to perform such tasks. In this research, a method for synthesizing insect pest training images in sticky paper trap images and enhancing the performance of insect pest convolutional neural network (CNN) classifier models through a generative adversarial network (GAN) is proposed. Through GAN, images were synthesized to balance and increase the number of training images even with limited data. The fidelity of the synthetic images was tested using t-SNE visualization. To evaluate the developed network, the minimum number of images required to apply GAN-based augmentation was determined. Furthermore, the performance of the insect pest classifier models without augmentation, with traditional augmentation, and with GAN-based augmentation were compared. Testing results show that the classifier models trained with images augmented using GAN achieved an F1-score of 0.95, outperforming the models trained with traditionally augmented images with an F1-score of 0.92. This research shows that by using GAN-based augmentation method, the insect pest CNN classification models can yield a better performance compared to classic augmentation method and reduce the effort needed for data collection. Furthermore, the proposed approach can be easily adapted to various agricultural applications using deep neural networks when image data is limited.

Toward Automatic Detection of Radiation-Induced Cerebral Microbleeds Using a 3D Deep Residual Network.

Cerebral microbleeds, which are small focal hemorrhages in the brain that are prevalent in many diseases, are gaining increasing attention due to their potential as surrogate markers of disease burden, clinical outcomes, and delayed effects of therapy. Manual detection is laborious and automatic detection and labeling of these lesions is challenging using traditional algorithms. Inspired by recent successes of deep convolutional neural networks in computer vision, we developed a 3D deep residual network that can distinguish true microbleeds from false positive mimics of a previously developed technique based on traditional algorithms. A dataset of 73 patients with radiation-induced cerebral microbleeds scanned at 7T with susceptibility-weighted imaging was used to train and evaluate our model. With the resulting network, we maintained 95% of the true microbleeds in 12 test patients and the average number of false positives was reduced by 89%, achieving a detection precision of 71.9%, higher than existing published methods. The likelihood score predicted by the network was also evaluated by comparing to a neuroradiologist's rating, and good correlation was observed.

Spiking neural networks for computer vision.

State-of-the-art computer vision systems use frame-based cameras that sample the visual scene as a series of high-resolution images. These are then processed using convolutional neural networks using neurons with continuous outputs. Biological vision systems use a quite different approach, where the eyes (cameras) sample the visual scene continuously, often with a non-uniform resolution, and generate neural spike events in response to changes in the scene. The resulting spatio-temporal patterns of events are then processed through networks of spiking neurons. Such event-based processing offers advantages in terms of focusing constrained resources on the most salient features of the perceived scene, and those advantages should also accrue to engineered vision systems based upon similar principles. Event-based vision sensors, and event-based processing exemplified by the SpiNNaker (Spiking Neural Network Architecture) machine, can be used to model the biological vision pathway at various levels of detail. Here we use this approach to explore structural synaptic plasticity as a possible mechanism whereby biological vision systems may learn the statistics of their inputs without supervision, pointing the way to engineered vision systems with similar online learning capabilities.

A Guide to Convolutional Neural Networks for Computer Vision

Computer vision has become increasingly important and effective in recent years due to its wide-ranging applications in areas as diverse as smart surveillance and monitoring, health and medicine, spor

Recent advances in convolutional neural networks

In the last few years, deep learning has led to very good performance on a variety of problems, such as visual recognition, speech recognition and natural language processing. Among different types of deep neural networks, convolutional neural networks have been most extensively studied. Leveraging on the rapid growth in the amount of the annotated data and the great improvements in the strengths of graphics processor units, the research on convolutional neural networks has been emerged swiftly and achieved state-of-the-art results on various tasks. In this paper, we provide a broad survey of the recent advances in convolutional neural networks. We detailize the improvements of CNN on different aspects, including layer design, activation function, loss function, regularization, optimization and fast computation. Besides, we also introduce various applications of convolutional neural networks in computer vision, speech and natural language processing.

Neural Network in Computer Vision for RoboCup Middle Size League

Robot World Cup Initiative (RoboCup) is a worldwide competition proposed to advance research in robotics and artificial intelligence. It has a league called RoboCup soccer devoted for soccer robots, which is a challenge because robots are mobile, fully autonomous, multi-agents, and they play on a dynamic environment. Moreover, robots must recognize the game entities, which is a crucial task during a game. A camera is usually used as an input system to recognize ball, opponents, soccer field, and so on. These elements may be recognized applying some tools of computational intelligence, for example an artificial neural network. This paper describes the application of an artificial neural network on middle size robotic football league, where a multilayer perceptron neural network is trained with the backpropagation algorithm, to classify elements on the image. Each output neuron represents an entity and its output value depends on the current entity that is present on the image. The results show that an artificial neural network successfully classified the entities. They were recognized even when similar color entities were present on the image.