Convolutional Neutral Network Research Articles

Deep convolution network for dense crowd counting

Estimating the total number of people in a crowded situation is a challenging task due to numerous occlusions and perspective changes existing in crowd images. To address this issue, the authors have proposed a new deep learning framework for accurate and efficient crowd counting here. Inspired by multi-column convolutional neural network (MCNN) and contextual pyramid convolutional neural network (CP-CNN), the authors use a combination of a two branches, convolutional neutral network (CNN) and transposed convolutional layers, to generate a high-quality density map. The two-branch CNN for feature extraction generates a density map that is only a quarter of size of the original image Then a set of transposed convolutional layers and convolutional layers are combined with the network to make up for the detail loss of the density map conducted by stacked pooling. Compared with MCNN and CP-CNN, the authors’ approach employs fewer branches and simpler architecture. Experimental result shows that their approach achieves MAE 80.7 and MSE 131.2 in ShanghaiTech PartA dataset, MAE 15.6 and MSE 26.8 in ShanghaiTech PartB dataset, and MAE Average 7.1 in WorldExpo'10 dataset.

A single-shot multi-level feature reused neural network for object detection

Recent years have witnessed the significant progress in object detection using deep convolutional neutral networks. However, there are few object detectors achieving high precision with low computational cost. In this paper, a novel and lightweight framework named multi-level feature reused detector (MFRDet) is proposed, which can reach a better accuracy than two-stage methods. It also can maintain comparable high efficiency of one-stage methods without employing very deep convolution neural networks as most modern detectors do. The proposed framework is suitable for reusing information included in deep and shallow feature maps, by which property the detection precision can be higher. For the Pascal VOC2007 test set trained with VOC 2007 and VOC 2012 training sets, the proposed MFRDet with the input size of 300 $$\times $$ 300 can achieve 80.7% mAP at the speed of 62.5 FPS. As for a high-resolution input version, MFRDet can obtain 82.0% mAP with the speed of 37.0 FPS using single Nvidia Tesla P100 GPU. The proposed framework shows the state-of-the-art mAP with high FPS, which is better than most of other modern object detectors.

Feedback2Code: A Deep Learning Approach to Identifying User-Feedback-Related Source Code Files

Users frequently raise feedback when using software products. Feedback from users regarding their experiences and expectations and software defects they found adds values to software maintenance and evolution — software managers collect user feedback and then dispatch feedback issues that developers (and/or maintainers) need to track and process. Feedback tracking is often supported by open source platforms and collaborative software systems. Meanwhile, there still exists a gap between feedback issues and source code: since user feedback is usually informal and arbitrary, engineers have to spend much effort on comprehending issues and identifying which source code files need to be improved or fixed. This paper introduces a deep learning approach, Feedback2Code , which facilitates identification of user-feedback-related source code files. The core idea is to (1) explore latent semantics of user feedback and source code using several deep learning techniques such as Multi-Layer Perceptron (MLP), Convolutional Neutral Network (CNN) and skip-gram and (2) establish a multi-correlation model to explore linkages between feedback issues and source code files. Given a feedback issue, the linkages then allow engineers to identify source code files that are highly relevant to the issue. We have implemented Feedback2Code and evaluated it against ChangeAdvisor (a state-of-the-art approach) on 24 open source projects. The evaluation results clearly show the strength of Feedback2Code : for 103793 feedback issues, Feedback2Code successfully established 101190 feedback-code linkages and achieved a precision that is [Formula: see text] higher than that of ChangeAdvisor . Feedback2Code also achieved an MRR and an MAP that are [Formula: see text] and [Formula: see text] higher than those of ChangeAdvisor , respectively. Furthermore, we also found that a Feedback2Code -trained model can be easily transferred, allowing feedback-code linkages to be established in new projects with a little history data.

A Review on Classifying Object through Machine Learning: International conference on Recent Trends in Artificial Intelligence, IOT, Smart Cities & Applications (ICAISC-2020)

Object detection has been a very important topic with the advancement of Computers vision systems. With the appearance of deep learning techniques, the accuracy for object detection has increased to their peak. Our project aims to include state-of-the-art technique for object detection to realize high accuracy with real-time performance. In this paper, we’ll provide a review of deep learning-based object detection framework, namely Convolutional Neutral Network (CNN) and advanced version of it namely, RCNN. Then we develop a system to detect multiple objects in a very single image. The network is trained on the foremost challenging publicly available data set (PASCAL VOC), on which object detection challenges is being conducted annually. The resulting system is fast and accurate

基于多尺度融合的卷积神经网络的杂草幼苗识别

基于多尺度融合的卷积神经网络的杂草幼苗识别

Effect of Multichannel Signal Sequence on Source Localization Using Convolutional Neural Network

In recent years, machine learning has become a promising data-driven method of source localization in underwater acoustics. Several algorithms have been developed by taking advantage of neural networks. This paper investigates the effect of multichannel signal sequence on the performance of source localization using a convolutional neutral network (CNN). In this paper, source localization is solved as a classification problem. The performances of different sequences are demonstrated to be quite different. For a specific CNN, it is revealed that the multichannel sequence affects source localization through influencing the complexity of range classification. The complexity can be reasonably reflected by the conspicuousness of signal differences between adjacent range categories. The two-dimensional (2D) Fourier spectrum of the signal differences provides an intuitive approach to describe the conspicuousness. The multichannel sequence that could induce greater spectral amplitudes has better localization performance in noisy environments.

A deeper look at the collective intelligence phenomenon

For the upcoming conference on Intelligence Studies at ICI 2020 in Bad Nauheim, Germany the focus of this issue of JISIB is on collective intelligence and foresight. The first two papers by Søilen and Almedia and Lesca deal with collective intelligence from an intelligence studies perspective. It may be said that the Internet itself is a gigantic collective intelligence effort, the largest in human history. Open source is a prerequisite for this system to work for everyone. The article by Černý et al. is on open source. All other contributions are on the connection between the Internet, software and intelligence. This issue consists of seven articles to compensate for two articles that were taken out by editors in the last issue. The first article by Søilen entitled “Making sense of the collective intelligence field: a review” is a historical review of the field of collective intelligence. The paper shows how collective intelligence is an interdisciplinary field and argues there is a flaw in the notion of “wisdom of crowds”. Collective intelligence can be understood in terms of social systems theory and as such this approach has been fruitful for the social sciences, although so far not very popular. It also bares relevance for the study of business and economics. The second article by Almeida and Lesca is entitled “Collective intelligence process to interpret weak signals and early warnings”. Early warning and the detection of weak signals is a vital topic for any intelligence organization. Two aspects are discussed in the paper, the importance of new technology and collective sense making or interpretation The third article by Shaikh and Singhal entitled “Study on the various intellectual property management strategies used and implemented by ICT firms for business intelligence” deals with intellectual property rights and patenting strategies. The authors identify a number of defensive and offensive IP strategies applied to ICT companies. The results have a bearing on patent acquisitions. The fourth article by Lamrhari et al. is entitled “Web intelligence for understanding customer satisfaction: application of Latent Dirichlet Allocation (LDA) and the Kano model”. Customer satisfaction today is mostly measured with data from the internet, using different business intelligence techniques. The Kano model is still valuablei,ii, but the way we gather information to assess the different levels in the model has changed. The authors use Latent Dirichlet Allocation to analyze the voice of customer (VOC) in online reviews. They suggest that BI techniques and a fuzzy-Kano model can enable companies to better understand their customers’ online reviews. The fifth article by Nahili et al. is entitled “A new corpus-based convolutional neutral network for big data text analysis”. Companies need efficient ways to analyze everything that is said about them on the internet (reviews, comments). The paper suggests a convolutional neural network (CNN) as it has been successfully used for text classification. IMDB movie reviews and Reuters datasets were used for the experiment. The sixth article by Černý et al. is entitled “Using open data and google search data for competitive intelligence analysis”. Taking the Czech antidepressant market as an example, the authors show how competitive intelligence can be obtained using Google Search data, Google Trend and other OSINT sources. The seventh article by Dadkhah et al. is entitled “The potential of business intelligence tools for expert findings”. The paper suggests a way for researchers to find experts using business intelligence tools. The same method may also be used by any business or person looking for experts on a specific topic. As always, we would above all like to thank the authors for their contributions to this issue of JISIB. Thanks to Dr. Allison Perrigo for reviewing English grammar and helping with layout design for all articles and to the Swedish Research Council for continuous financial support. We hope to see you all at the ICI 2020 on the 16-17 March, 2020. The deadline for the two-page abstract submission is March 1st, 2020.

A novel CNN-DDPG based AI-trader: Performance and roles in business operations

Artificial Intelligence (AI) is well-developed as a part of human life. In both financial markets and business operations, AI is getting more and more important. In this paper, we build a novel “Reinforcement Learning” (RL) framework based AI-trader. We adopt an actor-critic RL algorithm called “Deep Deterministic Policy Gradient” (DDPG) to find the optimal policy. Our proposed DDPG has two different convolutional neutral networks (CNNs) based function approximators. The proposed AI-trader’s performance is shown to outperform other methods with the use of real stock-index future data. We further discuss the generalization and implications of the proposed method for business operations.

Variational based Mixed Noise Removal with CNN Deep Learning Regularization.

In this paper, the traditional model based variational methods and deep learning based algorithms are naturally integrated to address mixed noise removal, specially for Gaussian mixture noise and Gaussian-impulse noise removal problem. To be different from single type noise (e.g. Gaussian) removal, it is a challenge problem to accurately discriminate noise types and levels for each pixel. We propose a variational method to iteratively estimate the noise parameters, and then the algorithm can automatically classify the noise according to the different statistical parameters. The proposed variational problem can be separated into regularization, synthesis, parameters estimation and noise classification four steps with the operator splitting scheme. Each step is related to an optimization subproblem. To enforce the regularization, the deep learning method is employed to learn the natural images prior. Compared with some model based regularizations, the CNN regularizer can significantly improve the quality of the restored images. Compared with some learning based methods, the synthesis step can produce better reconstructions by analyzing the types and levels of the recognized noise. In our method, the convolution neutral network (CNN) can be regarded as an operator which associated to a variational functional. From this viewpoint, the proposed method can be extended to many image reconstruction and inverse problems. Numerical experiments in the paper show that our method can achieve some state-of-the-art results for Gaussian mixture noise and Gaussian-impulse noise removal.

CNNDLP: A Method Based on Convolutional Autoencoder and Convolutional Neural Network with Adjacent Edge Attention for Predicting lncRNA-Disease Associations.

It is well known that the unusual expression of long non-coding RNAs (lncRNAs) is closely related to the physiological and pathological processes of diseases. Therefore, inferring the potential lncRNA–disease associations are helpful for understanding the molecular pathogenesis of diseases. Most previous methods have concentrated on the construction of shallow learning models in order to predict lncRNA-disease associations, while they have failed to deeply integrate heterogeneous multi-source data and to learn the low-dimensional feature representations from these data. We propose a method based on the convolutional neural network with the attention mechanism and convolutional autoencoder for predicting candidate disease-related lncRNAs, and refer to it as CNNDLP. CNNDLP integrates multiple kinds of data from heterogeneous sources, including the associations, interactions, and similarities related to the lncRNAs, diseases, and miRNAs. Two different embedding layers are established by combining the diverse biological premises about the cases that the lncRNAs are likely to associate with the diseases. We construct a novel prediction model based on the convolutional neural network with attention mechanism and convolutional autoencoder to learn the attention and the low-dimensional network representations of the lncRNA–disease pairs from the embedding layers. The different adjacent edges among the lncRNA, miRNA, and disease nodes have different contributions for association prediction. Hence, an attention mechanism at the adjacent edge level is established, and the left side of the model learns the attention representation of a pair of lncRNA and disease. A new type of lncRNA similarity and a new type of disease similarity are calculated by incorporating the topological structures of multiple bipartite networks. The low-dimensional network representation of the lncRNA-disease pairs is further learned by the autoencoder based convolutional neutral network on the right side of the model. The cross-validation experimental results confirm that CNNDLP has superior prediction performance compared to the state-of-the-art methods. Case studies on stomach cancer, breast cancer, and prostate cancer further show the ability of CNNDLP for discovering the potential disease lncRNAs.

Quantitative pathologic analysis of pulmonary nodules using three-dimensional computed tomography images based on latent Dirichlet allocation.

The main purpose of this paper is to quantificationally predict the pathologic characteristics of pulmonary nodules using a novel and effective computer assisted diagnosis (CADx) scheme based on latent Dirichlet allocation (LDA) model. To make use of LDA model, we propose a novel 3D rotation invariant LBP feature to construct image words through the K-means algorithm from 3D pulmonary nodule slices. A topic distribution for each pulmonary nodule can be acquired by well-trained LDA model, which was used for pathologic analysis based on rank-based statistical analysis. Using the LIDC/IDRI database, this study made experiments based on different parameters, including topic number and size of vocabulary. Experiments demonstrate that the performance of all the characteristics reached to accuracies of more than 80%. Especially, this study obtained an accuracy of 84.2% with the root mean square error (RMSE) of 1.068 on quantitative assessment of malignancy likelihood. Compared with the latest study of multi-task convolutional neutral network regression, the proposed method can obtain more accurate results of characteristic prediction of a pulmonary nodule.

A deep learning approach to detecting volcano deformation from satellite imagery using synthetic datasets

Satellites enable widespread, regional or global surveillance of volcanoes and can provide the first indication of volcanic unrest or eruption. Here we consider Interferometric Synthetic Aperture Radar (InSAR), which can be employed to detect surface deformation with a strong statistical link to eruption. Recent developments in technology as well as improved computational power have resulted in unprecedented quantities of monitoring data, which can no longer be inspected manually. The ability of machine learning to automatically identify signals of interest in these large InSAR datasets has already been demonstrated, but data-driven techniques, such as convolutional neutral networks (CNN) require balanced training datasets of positive and negative signals to effectively differentiate between real deformation and noise. As only a small proportion of volcanoes are deforming and atmospheric noise is ubiquitous, the use of machine learning for detecting volcanic unrest is more challenging than many other applications. In this paper, we address this problem using synthetic interferograms to train the AlexNet CNN. The synthetic interferograms are composed of 3 parts: 1) deformation patterns based on a Monte Carlo selection of parameters for analytic forward models, 2) stratified atmospheric effects derived from weather models and 3) turbulent atmospheric effects based on statistical simulations of correlated noise. The AlexNet architecture trained with synthetic data outperforms that trained using real interferograms alone, based on classification accuracy and positive predictive value (PPV). However, the models used to generate the synthetic signals are a simplification of the natural processes, so we retrain the CNN with a combined dataset consisting of synthetic models and selected real examples, achieving a final PPV of 82%. Although applying atmospheric corrections to the entire dataset is computationally expensive, it is relatively simple to apply them to the small subset of positive results. This further improves the detection performance without a significant increase in computational burden (PPV of 100%). Thus, we demonstrate that training with synthetic examples can improve the ability of CNNs to detect volcano deformation in satellite images, and propose an efficient workflow for the development of automated systems.

HERO: Human Emotions Recognition for Realizing Intelligent Internet of Things

Human emotions recognition (HERO) is considered as one of the important techniques for realizing the intelligent Internet of Things. The demand for a robust and precise facial expression recognition algorithm is urgent for the HERO. In this paper, we propose a deep recognition algorithm based on the ensemble deep learning model. The proposed algorithm consists of three sub-networks with different depths. Each sub-network is comprised of convolutional neutral networks and trained independently. The sub-network with more convolutional layers recognizes emotions by extracting local details such as the features of eyes and mouth, while the sub-network with less convolutional layers focuses on the macrostructure of the input image. The three sub-networks are assembled together to constitute the whole model. The experiment is based on the Kaggle facial expression recognition challenge database (FER2013), the Japanese female facial expression database, and the AffectNet database. The experimental results show that the proposed algorithm achieves a test accuracy of 71.91%, 96.44%, and 62.11% better than other competitors, and increases the test accuracy by approximately 2–3% than unique sub-networks.

Ship Classification and Detection Based on CNN Using GF-3 SAR Images

Ocean surveillance via high-resolution Synthetic Aperture Radar (SAR) imageries has been a hot issue because SAR is able to work in all-day and all-weather conditions. The launch of Chinese Gaofen-3 (GF-3) satellite has provided a large number of SAR imageries, making it possible to marine targets monitoring. However, it is difficult for traditional methods to extract effective features to classify and detect different types of marine targets in SAR images. This paper proposes a convolutional neutral network (CNN) model for marine target classification at patch level and an overall scheme for marine target detection in large-scale SAR images. First, eight types of marine targets in GF-3 SAR images are labelled based on feature analysis, building the datasets for further experiments. As for the classification task at patch level, a novel CNN model with six convolutional layers, three pooling layers, and two fully connected layers has been designed. With respect to the detection part, a Single Shot Multi-box Detector with a multi-resolution input (MR-SSD) is developed, which can extract more features at different resolution versions. In order to detect different targets in large-scale SAR images, a whole workflow including sea-land segmentation, cropping with overlapping, detection with MR-SSD model, coordinates mapping, and predicted boxes consolidation is developed. Experiments based on the GF-3 dataset demonstrate the merits of the proposed methods for marine target classification and detection.

Occlusion aware facial expression recognition using CNN with attention mechanism.

Facial expression recognition in the wild is challenging due to various un-constrained conditions. Although existing facial expression classifiers have been almost perfect on analyzing constrained frontal faces, they fail to perform well on partially occluded faces that are common in the wild. In this paper, we propose a Convolution Neutral Network with attention mechanism (ACNN) that can perceive the occlusion regions of the face and focus on the most discriminative unoccluded regions. ACNN is an end to end learning framework. It combines the multiple representations from facial regions of interest (ROIs). Each representation is weighed via a proposed Gate Unit that computes an adaptive weight from the region itself according to the unobstructed-ness and importance. Considering different RoIs, we introduce two versions of ACNN: patch based ACNN (pACNN) and global-local based ACNN (gACNN). pACNN only pays attention to local facial patches. gACNN integrates local representations at patch-level with global representation at image-level. The proposed ACNNs are evaluated on both real and synthetic occlusions, including a self-collected facial expression dataset with real-world occlusions (FED-RO), two largest in-the-wild facial expression datasets (RAF-DB and AffectNet) and their modifications with synthesized facial occlusions. Experimental results show that ACNNs improve the recognition accuracy on both the non-occluded faces and occluded faces. Visualization results demonstrate that, compared with the CNN without Gate Unit, ACNNs are capable of shifting the attention from the occluded patches to other related but unobstructed ones. ACNNs also outperform other state-of-the-art methods on several widely used in-the-lab facial expression datasets under the cross-dataset evaluation protocol.

Efficient automated detection of mitotic cells from breast histological images using deep convolution neutral network with wavelet decomposed patches

In medical practice, the mitotic cell count from histological images acts as a proliferative marker for cancer diagnosis. Therefore, an accurate method for detecting mitotic cells in histological images is essential for cancer screening. Manual evaluation of clinically relevant image features that might reflect mitotic cells in histological images is time-consuming and error prone, due to the heterogeneous physical characteristics of mitotic cells. Computer-assisted automated detection of mitotic cells could overcome these limitations of manual analysis and act as a useful tool for pathologists to make cancer diagnoses efficiently and accurately. Here, we propose a new approach for mitotic cell detection in breast histological images that uses a deep convolution neural network (CNN) with wavelet decomposed image patches. In this approach, raw image patches of 81 × 81 pixels are decomposed to patches of 21 × 21 pixels using Haar wavelet and subsequently used in developing a deep CNN model for automated detection of mitotic cells. The decomposition step reduces convolution time for mitotic cell detection relative to the use of raw image patches in conventional CNN models. The proposed deep network was tested using the MITOS (ICPR2012) and MITOS-ATYPIA-14 breast cancer histological datasets and shown to outperform existing algorithms for mitotic cell detection. Overall, our method improves the performance and reduces the computational burden of conventional deep CNN approaches for mitotic cell detection.

Deep learning for R -parity violating supersymmetry searches at the LHC

Supersymmetry with hadronic R-parity violation in which the lightest neutralino decays into three quarks is still weakly constrained. This work aims to further improve the current search for this scenario by the boosted decision tree method with additional information from jet substructure. In particular, we find a deep neural network turns out to perform well in characterizing the neutralino jet substructure. We first construct a Convolutional Neutral Network (CNN) which is capable of tagging the neutralino jet in any signal process by using the idea of jet image. When applied to pure jet samples, such a CNN outperforms the N-subjettiness variable by a factor of a few in tagging efficiency. Moreover, we find the method, which combines the CNN output and jet invariant mass, can perform better and is applicable to a wider range of neutralino mass than the CNN alone. Finally, the ATLAS search for the signal of gluino pair production with subsequent decay $\tilde{g} \to q q \tilde{\chi}^0_1 (\to q q q)$ is recasted as an application. In contrast to the pure sample, the heavy contamination among jets in this complex final state renders the discriminating powers of the CNN and N-subjettiness similar. By analyzing the jets substructure in events which pass the ATLAS cuts with our CNN method, the exclusion limit on gluino mass can be pushed up by $\sim200$ GeV for neutralino mass $\sim 100$ GeV.

Radio Galaxy Zoo:Claran– a deep learning classifier for radio morphologies

The upcoming next-generation large area radio continuum surveys can expect tens of millions of radio sources, rendering the traditional method for radio morphology classification through visual inspection unfeasible. We present ClaRAN - Classifying Radio sources Automatically with Neural networks - a proof-of-concept radio source morphology classifier based upon the Faster Region-based Convolutional Neutral Networks (Faster R-CNN) method. Specifically, we train and test ClaRAN on the FIRST and WISE images from the Radio Galaxy Zoo Data Release 1 catalogue. ClaRAN provides end users with automated identification of radio source morphology classifications from a simple input of a radio image and a counterpart infrared image of the same region. ClaRAN is the first open-source, end-to-end radio source morphology classifier that is capable of locating and associating discrete and extended components of radio sources in a fast (< 200 milliseconds per image) and accurate (>= 90 %) fashion. Future work will improve ClaRAN's relatively lower success rates in dealing with multi-source fields and will enable ClaRAN to identify sources on much larger fields without loss in classification accuracy.

Constraining the three-dimensional orbits of galaxies under ram pressure stripping with convolutional neural networks

Ram pressure stripping (RPS) of gas from disk galaxies has long been considered to play vital roles in galaxy evolution within groups and clusters. For a given density of intracluster medium (ICM) and a given velocity of a disk galaxy, RPS can be controlled by two angles (theta and phi) that define the angular relationship between the direction vector of the galaxy's three-dimensional (3D) motion within its host cluster and the galaxy's spin vector. We here propose a new method in which convolutional neutral networks (CNNs) are used to constrain theta and phi of disk galaxies under RPS. We first train a CNN by using ~10^5 synthesized images of gaseous distributions of the galaxies from numerous RPS models with different theta and phi. We then apply the trained CNN to a new test RPS model to predict theta and phi. The similarity between the correct and predicted theta and $\phi$ is measured by cosine similarity (cos-Theta) with cos-Theta =1 being perfectly accurate prediction. We show that the average cos-Theta among test models is ~0.95, which means that theta and phi can be constrained well by applying the CNN to the spatial distributions of their gas. This result suggests that if the ICM is in hydrostatic equilibrium (thus not moving), the 3D orbit of a disk galaxy within its host cluster can be constrained by the spatial distribution of the gas being stripped by RPS. We discuss the applications of the method to HI surveys such as WALLABY and SKA projects.

Application of Machine Learning to Classification of Volcanic Deformation in Routinely Generated InSAR Data

AbstractRecent improvements in the frequency, type, and availability of satellite images mean it is now feasible to routinely study volcanoes in remote and inaccessible regions, including those with no ground‐based monitoring. In particular, Interferometric Synthetic Aperture Radar data can detect surface deformation, which has a strong statistical link to eruption. However, the data set produced by the recently launched Sentinel‐1 satellite is too large to be manually analyzed on a global basis. In this study, we systematically process &gt;30,000 short‐term interferograms at over 900 volcanoes and apply machine learning algorithms to automatically detect volcanic ground deformation. We use a convolutional neutral network to classify interferometric fringes in wrapped interferograms with no atmospheric corrections. We employ a transfer learning strategy and test a range of pretrained networks, finding that AlexNet is best suited to this task. The positive results are checked by an expert and fed back for model updating. Following training with a combination of both positive and negative examples, this method reduced the number of interferograms to ∼100 which required further inspection, of which at least 39 are considered true positives. We demonstrate that machine learning can efficiently detect large, rapid deformation signals in wrapped interferograms, but further development is required to detect slow or small deformation patterns which do not generate multiple fringes in short duration interferograms. This study is the first to use machine learning approaches for detecting volcanic deformation in large data sets and demonstrates the potential of such techniques for developing alert systems based on satellite imagery.