Use Of Feature Information Research Articles

Asymmetric slack contrastive learning for full use of feature information in image translation

Recently, contrastive learning has been proven to be powerful in cross-domain feature learning and has been widely used in image translation tasks. However, these methods often overlook the differences between positive and negative samples regarding model optimization ability and treat them equally. This weakens the feature representation ability of the generative models. In this paper, we propose a novel image translation model based on asymmetric slack contrastive learning. We design a new contrastive loss asymmetrically by introducing a slack adjustment factor. Theoretical analysis shows that it can adaptively optimize and adjust according to different positive and negative samples and significantly improve optimization efficiency. In addition, to better preserve local structural relationships during image translation, we constructed a regional differential structural consistency correction block using differential vectors. Comparative experiments were conducted using seven existing methods on five datasets. The results indicate that our method can maintain structural consistency between cross-domain images at a deeper level. Furthermore, it is more effective in establishing real image-domain mapping relations, resulting in higher-quality images being generated.

AERMNet: Attention-enhanced relational memory network for medical image report generation

Background and objectivesThe automatic generation of medical image diagnostic reports can assist doctors in reducing their workload and improving the efficiency and accuracy of diagnosis. However, among the most existing report generation models, there are problems that the weak correlation between generated words and the lack of contextual information in the report generation process. MethodsTo address the above problems, we propose an Attention-Enhanced Relational Memory Network (AERMNet) model, where the relational memory module is continuously updated by the words generated in the previous time step to strengthen the correlation between words in generated medical image report. And the double LSTM with interaction module reduces the loss of context information and makes full use of feature information. Thus, more accurate disease information can be generated by AERMNet for medical image reports. ResultsExperimental results on four medical datasets Fetal heart (FH), Ultrasound, IU X-Ray and MIMIC-CXR, show that our proposed method outperforms some of the previous models with respect to language generation metrics (Cider improving by 2.4% on FH, Bleu1 improving by 2.4% on Ultrasound, Cider improving by 16.4% on IU X-Ray, Bleu2 improving by 9.7% on MIMIC-CXR). ConclusionsThis work promotes the development of medical image report generation and expands the prospects of computer-aided diagnosis applications. Our code is released at https://github.com/llttxx/AERMNET.

A Mixed Feature Extractor for Intelligent Fault Detection and Diagnosis of Tunnel Diode Circuit Systems

This paper presents a mixed feature extractor (MFE) for the fault detection and diagnosis of tunnel diode circuit systems described by Takagi-Sugeno (T-S) fuzzy model-based Markov jump systems (MJSs). A novel neural network model is constructed, which is composed of the 1-D convolutional neural network (CNN) and Transformer. In order to make full use of feature information, the 1-D CNN model is utilized to extract the local features, and the Transformer is established to obtain the global features. Then, the features taken from the MFE are concatenated and fed into a classification layer for fault detection and diagnosis. Finally, through experimental results, the proposed MFE is validated to be effective and outperform the commonly used diagnosis methods.

Brain Tumor Segmentation using Automatic 3D Multi-channel Feature Selection Convolutional Neural Network

Brain tumor segmentation is an important topic in medical image processing. Three-dimensional (3D) convolutional neural networks (CNN) make full use of feature information and have better segmentation performance. However, conventional 3D CNNs involve a significant number of parameters and has high hardware requirements, which is not conducive to clinical application. To solve these issues, this study proposes a lightweight highly-efficient 3D CNN for brain tumor magnetic resonance imaging segmentation. We first replaced the conventional CNN with improved depth-wise CNN to save computational resources. Next, we employed a multi-channel convolution approach using convolution kernels of different sizes to aggregate features from different receptive fields. Finally, we used a squeeze-and-excitation unit to automatically extract useful feature information in the network. Application of the proposed model to the BraTS2019 validation set resulted in mean Dice scores of 0.9053, 0.8373, and 0.7847, with regard to whole tumor, tumor core, and enhanced tumor, respectively. This was achieved with 2.01 M parameters and 24.68 G floating point operations, and data training used 2.75 GB of video memory. The differences between our Dice scores and those derived using the network developed by the winner of the BraTS2019 challenge were only 0.0041, 0.0174, and 0.0274. Furthermore, the number of parameters and floating point operations were 2.3 times and 8.1 times lower and data training used 4 times lower video memory than the winning network. This indicates that the proposed network offers a higher precision for brain tumor segmentation while reducing computational requirements and effectively lowering hardware requirements for clinical applications.

A Thermal Imaging Flame-Detection Model for Firefighting Robot Based on YOLOv4-F Model

Fire robots are an effective way to save lives from fire, but their limited detection accuracy has greatly hampered their practical applications in complicated fire conditions. This study therefore proposes an advanced thermal imaging flame detection model of YOLOv4-F based on YOLOv4-tiny. We replaced the Leaky ReLU activation function with the Mish activation function in the YOLOV4-tiny feature extraction network. A Spatial Pyramid Pooling (SPP) was also added to increase the receiving range of the feature extraction network. To improve the feature fusion efficiency between multi-scale feature layers, a Path Aggregation Network (PANet) was adopted to replace the YOLOv4-tiny Feature Pyramid Network (FPN) with full use of feature information; a high-quality dataset containing 14,757 thermal imaging flame images was built according to the PASCAL VOC 2007 dataset standard. The results show that, when compared to the YOLOv4-tiny, YOLOv5-s, and YOLOv7-tiny models, the average detection accuracy of the proposed YOLOv4-F model is 5.75% higher, the average mAP of the five IOU cases rises by 7.02%, and the average detection confidence of three scaled flames shows a 18.09% gain. The proposed YOLOV4-F meets the requirements of fire robots on real-time responses and accurate flame detection, offering an important tool to improve the performance of the current fire robots.

Interstitial fibrosis and tubular atrophy measurement via hierarchical extractions of kidney and atrophy regions with deep learning method

Background and objective:Interstitial fibrosis and tubular atrophy (IFTA) are the most critical characteristics of chronic kidney disease, and it is superior to glomerular pathology in predicting the decrease of glomerular filtration rate (eGFR) for chronic kidney disease (CKD). However, measurement of IFTA usually depends on manual operation, which is laborious and may lead to significant differences among pathologists. Therefore, the long-term focus of chronic kidney disease research has been on the glomerulus with relatively prominent morphological characteristics. There are few types of research on IFTA. Methods:We proposed an IFTA measurement method based on hierarchically extracting renal tissue, interstitial fibrosis and tubular atrophy regions. Our proposed method first extracts the whole renal tissue region (cortex and medulla) on WSIs by employing the Kidney Tissue Module and then recognizes the interstitial fibrosis and tubular atrophy regions via Interstitial fibrosis and Tubular atrophy Module. Finally, the proportion of the IFTA area to the entire renal tissue area is calculated to obtain the end-to-end output IFTA ratio. Results:Experimental results confirmed that the proposed method outperforms recently proposed approaches by comparing accuracy in segmenting IFTA area and thus achieved state-of-the-art performance. ITNet also achieved a faster-processing speed in calculating the proportion of IFTA than pathologists. Furthermore, the results showed that our proposed hierarchical extraction method could make full use of feature information, effectively combine the advantages of different scales, and segment IFTA with high accuracy and reasonable acceptability. Conclusion:Quantitative analysis of IFTA by the proposed method can reduce pathologists’ measurement time to diagnose and decrease the assessment difference between different pathologists. Therefore, the proposed method can assist pathologists in improving the efficiency of diagnosis and supplying consistent measurement accuracy.

Graph Convolution Network Recommender Rystem with Feature Embedding

Recommender system has become one of the most important parts of information services on the Internet today the model based on graph neural network has been proved to be a very effective method in collaborative filtering recommender system [1,15]. However, the recommender system based on graph network usually only uses the interactive relationship between users and items, and their feature information is often not used effectively. In this work, we combine the use of feature information with graph convolutional network to design a new recommendation algorithm, Graph Convolution Network recommender system with Feature Embedding, GCNFE. Tests on the Movie lens and Last-FM datasets show that the new model performs better than previous models in most cases.

EMRNet: End-to-End Electrical Model Restoration Network

The traditional method to improve the resolution in electromagnetic inversion is increasing the number of iterations, which displays poor non-linear mapping and strong non-uniqueness. To meet this challenge, a new strategy is proposed via reconstructing the geoelectric model for traitional inversion results through a deep neural networks (DNN). DNN possesses the advantage on establishing an uncertain mapping between low-resolution images and high-resolution target images. In order to recover the high-precision geoelectric model, we propose an end-to-end electromagnetic recovery network (EMRNet) with novel components to adequately utilize the geoelectric model data of traditional inversion. Specifically, EMRNet uses the codec structure from U-Net, whereby a cross-scale feature attention module (CSFA Block) is incorporated into the decoding process to make full use of feature information of different scales. The superiority of EMRNet are validated on both synthetic and measured data, The predicted geoelectric models of EMRNet are more consistent with the target from the aspects of resistivity values, overall structure, and resolution. In addition, the geoelectric model predicted by EMRNet agree well with the real geological background data and corresponding response data is closer to the measured data.

Use of features and geometry in leopard geckos (Eublepharis macularius)

All animals must move efficiently throughout their world. However, the mechanisms through which they accomplish this potentially vary among species. Previous work exploring the use of feature information and geometric information in movement through space has indicated that geometric information is commonly used and that some species sometimes also use feature information. Here, I investigated if a cold-blooded species, leopard geckos (Eublepharis macularius), would use geometric and/or feature information. In training, geckos learned to move to a correct corner within the box with a distinctive feature. In test when only geometric information was available, geckos chose either their assigned corner or its geometric opposite. In another test when feature information conflicted with geometric information, geckos did not use feature information and instead made choices consistent with using geometric information. This suggests geckos used geometric information preferentially to feature information in this experiment after both had been available throughout training when they were placed in conflict.

Generalized Variational Mode Decomposition: A Multiscale and Fixed-Frequency Decomposition Algorithm

To overcome the limitations of variational mode decomposition (VMD) algorithm that its frequency scales and spectrum positions cannot be flexibly adjusted to decompose signals as required, a generalized VMD (GVMD) was proposed. This article addresses the fundamental theory of GVMD. In order to highlight the local characteristics of the signal much more while considering its global data fidelity, a set of variational models is formed, where individual constrained optimization problem is constructed for each mode. The formed variational models are solved by the modified alternating direction method of multipliers approach, thus realizing the multiscale and fixed-frequency decomposition. To gain a deep insight into GVMD algorithm, its frequency band division manner is investigated. In essence, GVMD can be viewed as a bank of filters whose bandwidths and center frequencies can be flexibly adjusted by its parameters, i.e., scale parameters and prior center frequencies. The effectiveness of GVMD is verified on simulated and real signals. The preliminary results show that compared with state-of-the-art methods, GVMD can make full use of feature information to decompose original signals as desired into several narrowband modes or into several narrowband modes and a wideband mode, effectively obtaining the interested modes.

Classification of breast cancer histopathological images using interleaved DenseNet with SENet (IDSNet)

In this study, we proposed a novel convolutional neural network (CNN) architecture for classification of benign and malignant breast cancer (BC) in histological images. To improve the delivery and use of feature information, we chose the DenseNet as the basic building block and interleaved it with the squeeze-and-excitation (SENet) module. We conducted extensive experiments with the proposed framework by using the public domain BreakHis dataset and demonstrated that the proposed framework can produce significantly improved accuracy in BC classification, compared with the state-of-the-art CNN methods reported in the literature.

Change Detection Based on Feature Optimization in High Resolution Optical Image

Object-oriented change detection (CD) method can make full use of feature information of high resolution optical images. However, the feature information is redundancy in Object-oriented CD of high resolution optical image due to the fact that the image has multiple bands. Therefore, feature optimization is necessary to object-oriented CD. Aiming at the problem, a novel CD method based on feature optimization which combines improved locally linear embedding (ILLE) algorithm and object-oriented technology is proposed. Firstly, two temporal images are inverted into objects using multi-scale segmentation algorithm. Secondly, the spectral and texture features of the objects are extracted to construct the novel feature change vector. Thirdly, the improved LLE algorithm, which introduces the Geodesic distance metric, is designed to optimize the feature change vector. Finally, the CD result is obtained by FCM algorithm. Experiments construct on the real GF-1 images, and the results confirm the effectiveness of the proposed method.

A Novel Multi-Branch Channel Expansion Network for Garbage Image Classification

Due to the lack of data available for training, deep learning hardly performed well in the field of garbage image classification. We choose the TrashNet data set which is widely used in the field of garbage image classification, and try to overcome data deficiencies in this field by optimizing the network structure. In this article, it is found that the deeper network and short-circuit connection, which are generally accepted in the field of deep learning, will not work well on the TrashNet data set. By analyzing and modifying the network structure, we propose an effective method to improve the network performance on TrashNet data set. This method widens the network by expanding branches, and then uses add layers to realize the fusion of feature information. It can make full use of feature information at slight additional computational cost. Using this method to replace the core structure of the Xception network, the performance of the improved network has been improved greatly. Finally, the M-b Xception network proposed by us achieves 94.34% classification accuracy on the TrashNet data set, and has certain advantages over some state-of-the-art methods on multiple indicators. The python code can be download from https://github.com/scp19801980/Trash-classify-M_b-Xception.

Improved Active Sonar Tracking in Clutter Using Integrated Feature Data

Target tracking has traditionally focused on the use of kinematic information, such as bearing, range, position, or speed. In some cases, it is possible to extract additional target information that may be useful in the maintenance of reliable tracks, or the establishment of target type or identity. Several methods have been proposed for the use of feature information in target tracking. In this paper, we discuss and evaluate its use as a component of the cumulative likelihood ratio used for track management and contact assignment in a track-oriented multihypothesis tracker (MHT) using a sorted branch and bound approach to contact assignment. The active sonar sea trial data used in this evaluation, Clutter09, include feature values based on the perceptual capabilities of the human auditory system. Three measurement scenarios are considered. Comparisons between the baseline case, which used only kinematic information, and several feature-aided cases showed that the performance of the tracker improved noticeably when feature information was used. The MHT, which was working at a depth of six scans, was also able to tolerate seven times as many false contacts in the feature-aided cases as in the baseline cases. This corresponds to improved detection and tracking of targets that are weaker, or at greater range, or in a more cluttered environment. Results were also significantly better when compared with cases where the feature information was used to prefilter the contacts entering the tracker or to postfilter the tracks produced by the tracker.

FSNet: A Target Detection Algorithm Based on a Fusion Shared Network

Currently, in a target detection framework, a feature pyramid is widely used to capture differences resulting from the change in size of a detected object. Examples include the one-stage detector, the Deconvolutional Single-Shot Detector (DSSD), the RefineDet and the two-stage detector DetNet. Although these target detection frameworks with characteristic pyramid structure have achieved good results, they are limited by the facts that 1) they only use one feature layer of multi-scale information for prediction, and 2) they do not fully incorporate multi-scale inclusion of feature information. Therefore, based on Single-shot detector (SSD) architecture, we propose a novel one-stage target detector framework-fusion shared network (FSNet) which makes full use of feature information between multiple scale feature maps to more effectively detect objects of different scales. First, we introduce a fusion attention mechanism to fuse features between feature layers of different scales. Secondly, we combine this with multi-head structure fusion multi-layer features extracted as new features by a backbone network. Finally, we feed multi-layer fusion features into shared prediction module (shared PM). In the module, a new class feature pyramid for target detection is formed wherein the feature map used for prediction is composed of multiple levels of feature layers. In order to verify the effectiveness of this new FSNet target detection algorithm, we used the 320X320 input for FSNet, obtaining 80.7% mean accuracy (mAP) and 15 frames per second (Frame Per Second, FPS) in the PASCAL VOC 2007 test.

The Application of LCS and Information Entropy as a Novel Fusion Algorithm for Degradation Feature Extraction

Feature extraction is significant for the prognostics and health management (PHM) of hydraulic pumps. In this paper, a novel fusion algorithm is proposed based on local characteristic-scale decomposition (LCD), composite spectrum (LCS). and information entropy. To make full use of feature information, the LCS is proposed based on the modification of traditional composite spectral algorithm. LCS high-order power entropy and high-order singular entropy, which are relatively defined in Shannon entropy (SE) and Tsallis entropy (TE), are extracted as initial features. Furthermore, the method of feature fusion is presented to modify the features’ conciseness and to improve the performance. Results of the analysis in the experiment indicate that the proposed method is available, and the fused feature is effective in evaluating the pump degradation process.

Inductive Pairwise Ranking: Going Beyond the <i>n</i> log(<i>n</i>) Barrier

We study the problem of ranking a set of items from nonactively chosen pairwise preferences where each item has feature information with it. We propose and characterize a very broad class of preference matrices giving rise to the Feature Low Rank (FLR) model, which subsumes several models ranging from the classic Bradley–Terry–Luce (BTL) (Bradley and Terry 1952) and Thurstone (Thurstone 1927) models to the recently proposed blade-chest (Chen and Joachims 2016) and generic low-rank preference (Rajkumar and Agarwal 2016) models. We use the technique of matrix completion in the presence of side information to develop the Inductive Pairwise Ranking (IPR) algorithm that provably learns a good ranking under the FLR model, in a sample-efficient manner. In practice, through systematic synthetic simulations, we confirm our theoretical findings regarding improvements in the sample complexity due to the use of feature information. Moreover, on popular real-world preference learning datasets, with as less as 10% sampling of the pairwise comparisons, our method recovers a good ranking.

The Big Data Newsvendor: Practical Insights from Machine Learning

We investigate the data-driven newsvendor problem when one has n observations of p features related to the demand as well as historical demand data. We propose two approaches to finding the optimal order quantity in this new setting -- Machine Learning (ML) with and without regularization, and Kernel-weights Optimization (KO). We show that the resulting Big Data newsvendor problem can be solved by LP, MIP or QCQP programs under the ML approach, and by a simple sorting algorithm under the KO approach. We justify the use of feature information by showing that not including them yields inconsistent decisions, which translates to sub-optimal costs even with infinite amount of demand data. We then derive finite-sample performance bounds on the out-of-sample costs of the feature-based decisions, which shows (i) the Big Data regime, when over-fitting dominates finite-sample bias, is defined by p > O(n^{-1/(2 8/p)}\sqrt{\log{(n)}}), and (ii) both regularized ML and KO are effective methods to handle over-fitting. Finally, we apply the feature-based algorithms for nurse staffing in a hospital emergency room using a data set from a large UK teaching hospital and find that (i) the best KO and ML algorithms beat the best practice benchmark by 23% and 24% respectively in the out-of-sample cost with statistical significance at the 5% level, and (ii) the best KO algorithm is faster than the best ML algorithm by three orders of magnitude and the best practice benchmark by two orders of magnitude.