Pooling Method Research Articles

Efficient Training on Alzheimer's Disease Diagnosis with Learnable Weighted Pooling for 3D PET Brain Image Classification.

Three-dimensional convolutional neural networks (3D CNNs) have been widely applied to analyze Alzheimer's disease (AD) brain images for a better understanding of the disease progress or predicting the conversion from cognitively impaired (CU) or mild cognitive impairment status. It is well-known that training 3D-CNN is computationally expensive and with the potential of overfitting due to the small sample size available in the medical imaging field. Here we proposed a novel 3D-2D approach by converting a 3D brain image to a 2D fused image using a Learnable Weighted Pooling (LWP) method to improve efficient training and maintain comparable model performance. By the 3D-to-2D conversion, the proposed model can easily forward the fused 2D image through a pre-trained 2D model while achieving better performance over different 3D and 2D baselines. In the implementation, we chose to use ResNet34 for feature extraction as it outperformed other 2D CNN backbones. We further showed that the weights of the slices are location-dependent and the model performance relies on the 3D-to-2D fusion view, with the best outcomes from the coronal view. With the new approach, we were able to reduce 75% of the training time and increase the accuracy to 0.88, compared with conventional 3D CNNs, for classifying amyloid-beta PET imaging from the AD patients from the CU participants using the publicly available Alzheimer's Disease Neuroimaging Initiative dataset. The novel 3D-2D model may have profound implications for timely AD diagnosis in clinical settings in the future.

Point Cloud Deep Learning Network Based on Balanced Sampling and Hybrid Pooling

The automatic semantic segmentation of point cloud data is important for applications in the fields of machine vision, virtual reality, and smart cities. The processing capability of the point cloud segmentation method with PointNet++ as the baseline needs to be improved for extremely imbalanced point cloud scenes. To address this problem, in this study, we designed a weighted sampling method based on farthest point sampling (FPS), which adjusts the sampling weight value according to the loss value of the model to equalize the sampling process. We also introduced the relational learning of the neighborhood space of the sampling center point in the feature encoding process, where the feature importance is distinguished by using a self-attention model. Finally, the global-local features were aggregated and transmitted using the hybrid pooling method. The experimental results of the six-fold crossover experiment showed that on the S3DIS semantic segmentation dataset, the proposed network achieved 9.5% and 11.6% improvement in overall point-wise accuracy (OA) and mean of class-wise intersection over union (MIoU), respectively, compared with the baseline. On the Vaihingen dataset, the proposed network achieved 4.2% and 3.9% improvement in OA and MIoU, respectively, compared with the baseline. Compared with the segmentation results of other network models on public datasets, our algorithm achieves a good balance between OA and MIoU.

An improved method MSS-YOLOv5 for object detection with balancing speed-accuracy

For deep learning-based object detection, we present a superior network named MSS-YOLOv5, which not only considers the reliability in complex scenes but also promotes its timeliness to better adapt to practical scenarios. First of all, multi-scale information is integrated into different feature dimensions to improve the distinction and robustness of features. The design of the detectors increases the variety of detection boxes to accommodate a wider range of detected objects. Secondly, the pooling method is upgraded to obtain more detailed information. At last, we add the Angle cost and assign new weights to different loss functions to accelerate the convergence and improve the accuracy of network detection. In our network, we explore four variants MSS-YOLOv5s, MSS-YOLOv5m, MSS-YOLOv5x, and MSS-YOLOv5l. Experimental results of MSS-Yolov5s show that our technique improves mAP on the PASCAL VOC2007 and PASCAL 2012 datasets by 2.4% and 2.9%, respectively. Meanwhile, it maintains a fast inference speed. At the same time, the other three models have different degrees of performance improvement in terms of balancing speed and precision in challenging detection regions.

Macro-micro numerical simulation and experimental study of Mg-Gd-Y-Zn-Zr alloy fabricated by cold metal transfer wire arc additive manufacturing

The Mg-Gd-Y-Zn-Zr alloy components were fabricated by cold metal transfer (CMT) based wire arc additive manufacturing (WAAM). The heat transfer and equiaxed grain growth in the weld pool bonding zone of the alloy were investigated through macro-micro numerical simulation. The results shows, the morphology of the molten pool in the bonding area is ellipsoid with a major semi-axis of 13 mm and a semi-minor axis of 6 mm. The size of the molten pool increases with increasing temperature. With the increase in the number of layers of material added, the high temperature zone is enlarged continuously. This is due to the weakening of the heat dissipation of the substrate and the preheating effect of the first material added on the subsequent material added. The phase field model of equiaxed grain growth during the solidification process of the CMT arc welding pool, was established. The growth of equiaxed crystals in the bonding region temperature gradients of 4.9 °C/m and 3.6 °C/m, the propulsion velocity of 1.5 m/min and 2.3 m/min were simulated, respectively. The further away the molten pool from the keyhole center, the smaller the temperature gradient, the faster the dendrite growth speed, and the more vigorous the dendrite morphology. The unique layered processing method and Marangoni effect of the melting pool in the bonding zone of CMT-WAAM arc additive promote the β-Mg24(Gd,Y)5 and RE-rich phases of body-centered cubic structure. These precipitated phases are evenly distributed at the grain boundaries, it hinders the migration of grain boundarie st, which in turn inhibits grain growth.

AdaPool: Exponential Adaptive Pooling for Information-Retaining Downsampling.

Pooling layers are essential building blocks of convolutional neural networks (CNNs), to reduce computational overhead and increase the receptive fields of proceeding convolutional operations. Their goal is to produce downsampled volumes that closely resemble the input volume while, ideally, also being computationally and memory efficient. Meeting both these requirements remains a challenge. To this end, we propose an adaptive and exponentially weighted pooling method: adaPool. Our method learns a regional-specific fusion of two sets of pooling kernels that are based on the exponent of the Dice-Sørensen coefficient and the exponential maximum, respectively. AdaPool improves the preservation of detail on a range of tasks including image and video classification and object detection. A key property of adaPool is its bidirectional nature. In contrast to common pooling methods, the learned weights can also be used to upsample activation maps. We term this method adaUnPool. We evaluate adaUnPool on image and video super-resolution and frame interpolation. For benchmarking, we introduce Inter4K, a novel high-quality, high frame-rate video dataset. Our experiments demonstrate that adaPool systematically achieves better results across tasks and backbones, while introducing a minor additional computational and memory overhead.

The first serological detection of West Nile virus infection among residents living in northern Turkey.

West Nile virus (WNV) is transmitted by a mosquito-borne virus whose natural reservoir is birds. Humans and horses are considered accidental hosts. Even if the vast majority of WNV infections in humans have asymptomatic or mild disease settings, serious neurological disorders with lethal outcomes can also be observed in around 1% of the cases. We aimed to serologically investigate the presence of WNV in humans living in Black sea of Turkey, and to obtain epidemiological data that will contribute to the implementation of public health policies to control and prevent potentially other life-threatening arboviral infections. In the current study, a total of 416 human sera were collected from native patients of Samsun and its boroughs attending Samsun Training and Research Hospital; these sera were tested for WNV with pooling method, using anti-IgM and IgG ELISA commercial kits. All pools that were found positive for both IgM and IgG were individually retested for the detection of positive WNV sera. After that, all positive samples were tested using real-time PCR to detect the presence of WNV-RNA particles. Total seropositivity rates of WNV in terms of IgM and IgG were found as 0.96% and 0.72%, respectively. No presence of WNV-RNA could be detected in positive samples. According to the data, further studies should be conducted to better understand the epidemiological dynamics of WNV in Turkey. It is recommended that other antigenically related flaviviruses which can give cross-reaction with WNV should also be investigated.

Dynamic Allocation Method of Incentive Pool for Financial Management Teaching Innovation Team Based on Data Mining

Dynamic Allocation Method of Incentive Pool for Financial Management Teaching Innovation Team Based on Data Mining

Deep Learning Detection Method of All-position Welding Pool

Deep Learning Detection Method of All-position Welding Pool

Heterogeneous Graph Convolutional Neural Network via Hodge-Laplacian for Brain Functional Data.

This study proposes a novel heterogeneous graph convolutional neural network (HGCNN) to handle complex brain fMRI data at regional and across-region levels. We introduce a generic formulation of spectral filters on heterogeneous graphs by introducing the k-th Hodge-Laplacian (HL) operator. In particular, we propose Laguerre polynomial approximations of HL spectral filters and prove that their spatial localization on graphs is related to the polynomial order. Furthermore, based on the bijection property of boundary operators on simplex graphs, we introduce a generic topological graph pooling (TGPool) method that can be used at any dimensional simplices. This study designs HL-node, HL-edge, and HL-HGCNN neural networks to learn signal representation at a graph node, edge levels, and both, respectively. Our experiments employ fMRI from the Adolescent Brain Cognitive Development (ABCD; n=7693) to predict general intelligence. Our results demonstrate the advantage of the HL-edge network over the HL-node network when functional brain connectivity is considered as features. The HL-HGCNN outperforms the state-of-the-art graph neural networks (GNNs) approaches, such as GAT, BrainGNN, dGCN, BrainNetCNN, and Hypergraph NN. The functional connectivity features learned from the HL-HGCNN are meaningful in interpreting neural circuits related to general intelligence.

Temporal Attention-Pyramid Pooling for Temporal Action Detection

Temporal action detection is a challenging task in video understanding, which is usually divided into two stages: proposal generation and classification. Learning proposal features is a crucial step for both stages. However, most methods ignore temporal information of proposals and consider background and action frames in proposals equally, leading to poor proposal features. In this paper, we propose a novel Temporal Attention-Pyramid Pooling (TAPP) method to learn proposal features of arbitrary length action proposals. The TAPP method exploits the attention mechanism to focus on the discriminative part of proposals, suppressing background influence on proposal features. It constructs a temporal pyramid structure to convert arbitrary length proposal feature sequences to multiple fixed-length sequences while retaining the temporal information. In the TAPP method, we design a multi-scale temporal function and apply it to the temporal pyramid to generate final proposal features. Based on the TAPP method, we construct a temporal action proposal generation model and an action proposal classification model, and then we perform extensive experiments on two mainstream temporal action detection datasets for the temporal action proposal and temporal action detection tasks to verify our models. On the THUMOS'14 dataset, our models based on the TAPP significantly outperform the previous state-of-the-art methods for both tasks.

Evaluation of Brucella pericarditis cases by pooled analysis method

Evaluation of Brucella pericarditis cases by pooled analysis method

Environmental Sound Classification With Low-Complexity Convolutional Neural Network Empowered by Sparse Salient Region Pooling

Environmental Sound Classification (ESC) is an important field in a broad range of applications, such as smart cities, audio surveillance, and health care. Recently, Convolutional Neural Networks (CNNs) have taken the lead from traditional approaches and have produced promising results. However, the achieved improvements are often accompanied by increasing depth, complexity, and size of the network, which prevents their usage in many practical applications. In this work, our goal is to empower a small-size low-complexity CNN model to achieve superior performance. To this end, we concentrate on the importance of global pooling technique, which is less investigated in ESC. In most previous works, models utilize global average pooling layer which does not consider regional saliency, and thus weakens the salient time-frequency regions contributions to the classification, and also to the training of convolutional kernels. We propose a novel global pooling method, called Sparse Salient Region Pooling (SSRP), which computes the channel descriptors using a sparse subset of features, and guides the model to effectively learn from the more salient time-frequency regions. Experimental results demonstrate that the proposed model with only 700K parameters yields accuracies of 86.7% on ESC-50 and 94.8% on ESC-10, which are comparable to that of the state-of-the-art methods. Compared to the baseline model, our model achieves absolute improvement of 21.8% in accuracy on ESC-50, with 98% smaller model size. Our visual analyses show that SSRP intensifies the responses of low-energy regions such that they contribute even more than high-energy regions to the classification of specific sound classes.

Unknown Binary Protocol Recognition Algorithm Based on One Class of Classification and One‐Dimensional CNN

In order to solve the problem that there are a large number of unknown protocols on the network, which affect the network management and network security in varying degrees, an unknown binary protocol identification method is proposed. On the premise that the center cluster of unknown protocols is obtained by the clustering algorithm, the unknown protocols for network traffic are identified by combining one‐class of classification with one‐dimensional CNN classification technology. At first, a class of classification algorithm is used to select the unknown protocols, and then, the labeled protocol data obtained by clustering are used to train the one‐dimensional CNN model, and the classified binary protocol packets are directly used as the input of the one‐dimensional convolution neural network. After the classification of the CNN model, the unknown protocols are finally identified. The experimental results show that the proposed classification and recognition method is better than the traditional CNN and SVM algorithms, and the maximum frequency pooling is better than the traditional pooling method.

Результати вирощування цьоголіток судака (Sander lucioperca Linnaeus, 1758) із застосуванням басейнової технології

Purpose.To investigate the peculiarities of pikeperch fingerlings (Sander lucioperca Linnaeus, 1758) cultivated with the use of pool technology and different stoking densities, also to provide a comprehensive fishery-biological assessment of the produced fish. Methodology. The material for the study were juvenile and young-of-the-year pikeperch. They were cultivated in 17.3 m3 pools (in the first series of the experiment) and 2.16 m3 pools (in the second series of the experiment). According to the physiological needs of pikeperch at the studied stage of ontogenesis, its diet consisted of salt brine shrimp (Artemia salina Linnaeus, 1758) nauplii and combined feed. They were fed according to a scheme with a sequential change of 3 stages: 1 – feeding exclusively with brine shrimp nauplii; 2 – combined feeding (brine shrimp nauplii and combined feed were mixed); 3 – feeding exclusively with combined feed. Also, 3 options of stocking density were used at each of the above-mentioned growing stages. Namely: at the 1st stage ‒ 40, 60 and 80 thousand individuals/m3; in the II stage ‒ 10, 15 and 20 thousand individuals/m3, and in the III stage ‒ 5, 7, 10 thousand individuals/m3. During the entire growing process, according to the schedule, test catches were carried out, during which weighing was carried out, based on the results of which absolute and relative weight gains were calculated. Statistical processing of the received data was carried out using a package of standard Microsoft Office software. Findings. The study established that the highest growth rates were found in fish that were grown at the stocking density: at first 40 thousand individuals/m3 (I stage), and then 10 thousand individuals/m3 (II stage) and 5 (III stage) individuals/m3 after that — experimental group No 1. The individual weights in the fish from this experimental group were 1.7–1.8 g at the first stage of rearing, and at the second stage of rearing, the intensity of weight accumulation in them increased almost twice (2.4–2.2 g). The survival rate of fish from the experimental group No 1 was 37.5–41.0% at the I stage of rearing and 74.0–78.0% at the II stage. Under such scheme (feeding and stocking density as in experimental group No 1) of pikeperch juveniles and young-of-the-year cultivation in pools, the fish productivity of the latter ranged from 21 to 35.8 kg/m3 (depending on the initial size of the pools for rearing). Originality.For the first time in the conditions of Ukraine, a comparative analysis of young-of-the-pikeperch cultivation using intensive technology was carried out, and practical recommendations were given to improve the existing technology. Practical Value.The use of the pool method of young-of-the-pikeperch cultivation allows expanding the existing technologies of aquaculture for this fish species and increase the market of its additional products. Keywords: pikeperch (Sander lucioperca Linnaeus, 1758), juveniles, fingerlings, pool technology of aquaculture, planting densities, weight, survival.

Multi-granularity graph pooling for video-based person re-identification

The video-based person re-identification (ReID) aims to identify the given pedestrian video sequence across multiple non-overlapping cameras. To aggregate the temporal and spatial features of the video samples, the graph neural networks (GNNs) are introduced. However, existing graph-based models, like STGCN, perform the mean/max pooling on node features to obtain the graph representation, which neglect the graph topology and node importance. In this paper, we propose the graph pooling network (GPNet) to learn the multi-granularity graph representation for the video retrieval, where the graph pooling layer is implemented to downsample the graph. We construct a multi-granular graph by using node features learned from backbone, then implement multiple graph convolutional layers to perform the spatial and temporal aggregation on nodes. To downsample the graph, we propose a multi-head full attention graph pooling (MHFAPool) layer, which integrates the advantages of existing node clustering and node selection pooling methods. Specifically, MHFAPool first learns a full attention matrix for each pooled node, then obtains the principal eigenvector of the attention matrix via the power iteration algorithm, finally takes the softmax of the principal eigenvector as the aggregation coefficients. Extensive experiments demonstrate that our GPNet achieves the competitive results on four widely-used datasets, i.e., MARS, DukeMTMC-VideoReID, iLIDS-VID and PRID-2011.

Appropriately estimating the standardized average treatment effect with missing data: A simulation and primer.

Reporting standardized effects in randomized treatment studies aids interpretation and facilitates future meta-analyses and policy considerations. However, when outcome data are missing, achieving an unbiased, accurate estimate of the standardized average treatment effect, sATE, can pose challenges even for those with general knowledge of missing data handling, given that the sATE is a ratio of a mean difference to a (within-group) standard deviation. Under both homogeneity and heterogeneity of variance, a Monte Carlo simulation study was conducted to compare missing data handling strategies in terms of bias and accuracy in the sATE, under specific missingness patterns plausible for randomized pretest posttest studies. Within two broad missing data handling approaches, maximum likelihood and multiple imputation, modeling choices were thoroughly investigated including the analysis model, variance estimator, imputation algorithm, and method of pooling results across imputed datasets. Results demonstrated that although the sATE can be estimated with little bias using either maximum likelihood or multiple imputation, particular attention should be paid to the model and variance estimator, especially at smaller sample sizes (i.e., N = 50). Differences in accuracy were driven by differences in bias. To improve estimation of the sATE in practice, recommendations and a software demonstration are provided. Moreover, a pedagogical explanation of the causes of bias, described separately for the numerator and denominator of the sATE is provided, demonstrating visually how and why bias occurs with certain methods.

CGUN-2A: Deep Graph Convolutional Network via Contrastive Learning for Large-Scale Zero-Shot Image Classification

Taxonomy illustrates that natural creatures can be classified with a hierarchy. The connections between species are explicit and objective and can be organized into a knowledge graph (KG). It is a challenging task to mine features of known categories from KG and to reason on unknown categories. Graph Convolutional Network (GCN) has recently been viewed as a potential approach to zero-shot learning. GCN enables knowledge transfer by sharing the statistical strength of nodes in the graph. More layers of graph convolution are stacked in order to aggregate the hierarchical information in the KG. However, the Laplacian over-smoothing problem will be severe as the number of GCN layers deepens, which leads the features between nodes toward a tendency to be similar and degrade the performance of zero-shot image classification tasks. We consider two parts to mitigate the Laplacian over-smoothing problem, namely reducing the invalid node aggregation and improving the discriminability among nodes in the deep graph network. We propose a top-k graph pooling method based on the self-attention mechanism to control specific node aggregation, and we introduce a dual structural symmetric knowledge graph additionally to enhance the representation of nodes in the latent space. Finally, we apply these new concepts to the recently widely used contrastive learning framework and propose a novel Contrastive Graph U-Net with two Attention-based graph pooling (Att-gPool) layers, CGUN-2A, which explicitly alleviates the Laplacian over-smoothing problem. To evaluate the performance of the method on complex real-world scenes, we test it on the large-scale zero-shot image classification dataset. Extensive experiments show the positive effect of allowing nodes to perform specific aggregation, as well as homogeneous graph comparison, in our deep graph network. We show how it significantly boosts zero-shot image classification performance. The Hit@1 accuracy is 17.5% relatively higher than the baseline model on the ImageNet21K dataset.

A Parallel Multi-Modal Factorized Bilinear Pooling Fusion Method Based on the Semi-Tensor Product for Emotion Recognition.

Multi-modal fusion can exploit complementary information from various modalities and improve the accuracy of prediction or classification tasks. In this paper, we propose a parallel, multi-modal, factorized, bilinear pooling method based on a semi-tensor product (STP) for information fusion in emotion recognition. Initially, we apply the STP to factorize a high-dimensional weight matrix into two low-rank factor matrices without dimension matching constraints. Next, we project the multi-modal features to the low-dimensional matrices and perform multiplication based on the STP to capture the rich interactions between the features. Finally, we utilize an STP-pooling method to reduce the dimensionality to get the final features. This method can achieve the information fusion between modalities of different scales and dimensions and avoids data redundancy due to dimension matching. Experimental verification of the proposed method on the emotion-recognition task using the IEMOCAP and CMU-MOSI datasets showed a significant reduction in storage space and recognition time. The results also validate that the proposed method improves the performance and reduces both the training time and the number of parameters.

Anomaly Detection Method in Railway Using Signal Processing and Deep Learning

In this paper, anomaly detection of wheel flats based on signal processing and deep learning techniques is analyzed. Wheel flats mostly affect running stability and ride comfort. Currently, domestic railway companies visually inspect wheel flats one by one with their eyes after railway vehicles enter the railway depots for maintenance. Therefore, CBM (Condition-Based Maintenance) is required for wheel flats resolution. Anomaly detection for wheel flat signals of railway vehicles using Order analysis and STFT (Short Time Fourier Transform) is studied in this paper. In the case of railway vehicles, it is not easy to obtain actual failure data through running vehicles in a university laboratory due to safety and cost issues. Therefore, vibration-induced acceleration was obtained using a multibody dynamics simulation software, SIMPACK. This method is also proved in the other paper by rig tests. In addition, since the noise signal was not included in the simulated vibration, the noise signal obtained from the Seoul Metro Subway Line 7 vehicle was overlapped with the simulated one. Finally, to improve the performance of both detection rate and real-time of characteristics based on existing LeNet-5 architectures, spectrogram images transformed from time domain data were proceeded with the LeNet deep learning model modified with the pooling method and activation function. As a result, it is validated that the method using the spectrogram with a deep learning approach yields higher accuracy than the time domain data.

Wetland mitigation functions on hydrological droughts: From drought characteristics to propagation of meteorological droughts to hydrological droughts

Wetlands have been singled out as a potential nature-based solution for improving the resilience and reducing the risks of hydrometeorological extremes. However, whether and to what extent wetlands can affect hydrological droughts is not well understood. To fill this gap, we proposed a general framework to discern the effect of wetlands on: (i) the characteristics (duration, severity, development and recovery processes) of hydrological droughts, and (ii) the propagation of meteorological droughts to hydrological droughts. First, hydrological modeling was conducted with a spatially-explicit model integrated with wetland modules. Then, the run theory and pooling method were selected to recognize hydrological drought events and identify their characteristics. Further, the Pearson correlation coefficient, temporal shift method and cross wavelet transform were used to explore the propagation processes. Finally, the characteristics and propagation processes were compared to quantify wetland mitigation services on hydrological droughts. To validate the proposed framework, two river basins from China and Canada (the Gan River Bain and the Nelson River Bain), with distinct land cover, were chosen to perform hydrological modeling and quantify wetland effects. The results indicate that wetlands mainly contribute to alleviating hydrological droughts by decelerating the development process, accelerating the recovery, shortening the duration, and reducing the severity of the hydrological drought events. However, the effects are variable as they may have weak impacts and even worsen drought conditions. Wetlands can extend drought propagation time and weaken the transition of meteorological to hydrological droughts. The likelihood of hydrological drought formation due to meteorological decreased by 19% and 18% respectively, for the Gan River Bain and Nelson River Bain, thanks to the mitigation services of wetlands. These findings highlight the drought-mitigation roles of wetlands and the proposed modelling framework has the potential to be useful in beneficial in assisting basin management on drought risks in the context of climate change mitigation.