Search

Objectives: Convolutional neural network (CNN) was widely used in landslide susceptibility assessment because of its powerful feature extraction capability. However, with the demand for scene diversification and high accuracy, the algorithm of CNN was constantly improved. The practice of improving accuracy by deepening the network hierarchy or combining other models tended to increase the number of model parameters and the amount of calculation greatly, which led to the difficulty of model training or the overfitting of results, thus limiting its practical application. We want to solve the above problems from the model structure. of feature maps, one-dimensional CNN（1D-CNN） and two-dimensional CNN（2D-CNN） were connected to maintain network depth, limit model parameters, and reduce computation. The parameters sharing of multidimensional convolution kernels was used to capture the deep coupling features of different dimensions and different landslide factors, so as to make full use of features and avoid overfitting. Taking Sedongpu gully in Xizang,China as the experimental area, this paper selected 11 kinds of landslide influencing factors to analyze the landslide susceptibility. Results: The results show that the multi-dimensional CNN coupling structure had the same training efficiency as the shallow 2D-CNN with fewer parameters due to the reduced computational effort. While compared with the deep 2D-CNN with the approximate number of parameters, the training time of the proposed method was significantly reduced and the training cost was lower. In addition, the coupled model had a stronger feature learning ability than the independent 1D-CNN and 2D-CNN, and therefore obtained higher model accuracy. Under each confusion matrix metric of the testing data, the coupled model received higher scores, and thus obtained more reliable landslide susceptibility assessment results. Conclusions: The multidimensional CNN coupling model proposed in this paper is a reliable method applicable to landslide susceptibility assessment. This study provides new theoretical guidance and technical support for further landslide hazard monitoring and prevention [ABSTRACT FROM AUTHOR]

Aiming at the problems of low recognition accuracy and large calculation amount of age estimation, this paper proposes a staged single aggregation backbone network VoVNetv4. In order to improve the feature extraction ability of the network, the Regional One-Shot Aggregation (ROSA) module is proposed in the network for regional encapsulation of the feature module. The classifier adopts the COnsistent RAnk Logits (CORAL) framework, which can solve the problem of inconsistency of classifiers in ordered regression CNNs for multiple binary classification tasks. This paper conducts experimental verification on the Asian Face Age Dataset (AFAD) and compares it with other research methods. The experimental results show that VoVNetv4 can effectively improve the feature extraction and detection ability of the network and improve the detection efficiency of age estimation. [ABSTRACT FROM AUTHOR]

To solve the problem of inaccurate description and unreliable relation inference in group activity recognition, this paper focused on constructing a scene relationship graph for three aspects: individual, group, and scene, and proposed a scene relationship graph network（SRGN） for group activity recognition. This method included a feature extraction module, a scene relation graph inference module, and a classification module. The feature extraction module extracted individual features, group features, and scene features by convolutional neural network. To fully explore the impact of scene on individual and group descriptions, the scene relation graph inference module learnt individual features and group features by building individual-scene and group-scene relationship graphs in a two-branch framework. Scene graph inference took into account the influence of individual on group and introduced a cross-branch module. It used the classification module to classify individual features and group features for prediction. The experimental results show that the group recognition accuracy of the proposed method on volleyball and collective activity data sets is improved by 1. 1% and 0. 5%,respectively. It verifies the validity of the scene graph in describing individual feature and group feature. [ABSTRACT FROM AUTHOR]

The road network system is the core foundation of a city. Extracting road information from remote sensing images has become an important research direction in the current traffic information industry. The efficient residual factorized convolutional neural network (ERFNet) is a residual convolutional neural network with good application value in the field of biological information, but it has a weak effect on urban road network extraction. To solve this problem, we developed a road network extraction method for remote sensing images by using an improved ERFNet network. First, the design of the network structure is based on an ERFNet; we added the DoubleConv module and increased the number of dilated convolution operations to build the road network extraction model. Second, in the training process, the strategy of dynamically setting the learning rate is adopted and combined with batch normalization and dropout methods to avoid overfitting and enhance the generalization ability of the model. Finally, the morphological filtering method is used to eliminate the image noise, and the ultimate extraction result of the road network is obtained. The experimental results show that the method proposed in this paper has an average F1 score of 93.37% for five test images, which is superior to the ERFNet (91.31%) and U-net (87.34%). The average value of IoU is 77.35%, which is also better than ERFNet (71.08%) and U-net (65.64%). [ABSTRACT FROM AUTHOR]

To solve the problem of combustion state recognition in solid oxide fuel cell (SOFC) combustor, this paper proposed a combustion state recognition method based on attention mechanism and image feature pyramid. The method adopted the adaptive gamma correction with weighting distribution (AGCWD) to standardize the input images. It combined two single full connections with 1×1 convolution to replace the squeeze and excitation structure, and proposed a hybrid attention structure combined with spatial attention structure to enhance the ability of feature extraction. To provide multi-scale information communication capability, it constructed the multi-scale bidirectional fusion pyramid by means of bidirectional computation and multi-scale fusion. The experimental results show that the proposed method reaches 99.22% accuracy under the premise of 3.98 M parameters and 397 M floating point operations (FLOPs), and effectively identifies the combustion state in SOFC combustor. [ABSTRACT FROM AUTHOR]

In image retrieval based on deep hashing, to solve the problems of low feature extraction efficiency in convolutional neural networks (CNN ) and underutilization of feature correlation, this paper proposed a novel method combining sparse difference network and multi-supervised hashing( SDNMSH ), and used it for efficient image retrieval. SDNMSH took pairs of images as training inputs, and guided hash codes learning through an elaborately designed sparse difference convolutional neural network and a supervised hash function. The sparse difference convolutional layer and the vanilla convolutional layer composed the sparse difference convolutional neural network. The sparse difference convolutional layer could quickly extract rich feature information, to achieve efficient feature extraction of the entire network. At the same time, in order to make full use of the pairwise correlation of semantic information and features, so as to promote the feature information extracted by the network to be more effectively transformed into discriminative hash codes, and then to achieve efficient image retrieval by using SDNMSH, this paper adopted a multi-supervised hash( MSH) function and designed an objective function for this purpose. Extensive experimental results on three widely used datasets MNIST,CIFAR-10 and NUS-WIDE show that SDNMSH achieves better retrieval performance, compared with the state-of-the-arts. [ABSTRACT FROM AUTHOR]

Most of the existing steganalysis methods have weak generalization ability and cannot effectively detect unknown steganalysis algorithms, which makes the accuracy of their classification greatly reduced in the practical application process. To solve this problem, this paper proposed an image steganalysis method based on group convolution and snapshot ensembling（Snapshot Ensembling steganalysis network, SENet）. Firstly, residual convolution block and group convolution block are used to extract the features of the image. Secondly, the model with the best performance is obtained as the snapshot model in each training period. Finally, the selected snapshot models are integrated to classify the images. This method uses the techniques of group convolution and snapshot ensembling to avoid the high training cost of traditional integration methods and the limited generalization ability of a single classifier. Experimental results show that this method can improve the accuracy of steganographic analysis model, and can effectively classify when the training set and test set are mismatched, and has high model generalization ability. [ABSTRACT FROM AUTHOR]

In order to solve the problem that the Transformer model cannot apply to mobile devices with limited computational resources due to its huge number of parameters and computational complexity, this paper proposed a mobile-friendly lightweight image recognition network called FuseNet. FuseNet utilized convolutional neural network to extract local feature information and self-attention to excel in modeling global information, and it integrated the features of both local and global representations into a single feature extraction module, which efficiently combined the advantages of the two different structures to achieve a high accuracy with a small model size. Experiments demonstrate that FuseNet with different model sizes can achieve good performance without using pre-training and it can well apply to mobile devices. FuseNet-B achieves 80. 5% accuracy with 14. 8 M parameters on ImageNet-1 K dataset, the performance exceeds the same volume of Transformer models and convolutional neural networks. [ABSTRACT FROM AUTHOR]

In the field of CXR image classification, current studies mostly use convolutional neural network and transfer learning technology. But in the process of rapidly constructing a network without considering the specificity of the CXR image. To solve the above problem, this paper proposed a novel XDense-RC-net. Based on the DenseNet model, this method improved the capability of feature extraction and feature fusion by introducing a new spatial attention mechanism in the densely connected layer, Optimized Transition module of DenseNet by using two different pooling strategies to enhance the noise immunity of the model. XDense-RC-net used the chest X-ray14 (multilabel and 14-class) and COVIDx (single label and 3-class) datasets for validating. In the multi-label classification experiments, the average AUC score reached 0.854, which is 0.109 higher than the benchmark method. In the single-label classification experiments, the average accuracy reached 96.75%, which is 7.75% higher than the baseline model. The results show that XDense-RC-net improves the accuracy of CXR image classification and can generalize to multilabel and single-label classification tasks. [ABSTRACT FROM AUTHOR]

The depth estimation algorithm based on the convolutional neural network has many limitations and defects by constructing matching cost volume to calculate the disparity: using a limited disparity range, the authentic disparity beyond the predetermined range can not be acquired; Besides, the matching process lacks constraints on occlusion and matching uniqueness; Also, as a local feature extractor, a convolutional neural network lacks the ability of global context information perception. Aiming at the problems in the matching method of constructing matching cost volume, we propose a disparity prediction algorithm based on Transformer, which specifically comprises the Swin-SPP module for feature extraction based on Swin Transformer, Transformer disparity matching network based on self-attention and cross-attention mechanism, and occlusion prediction sub-network. In addition, we propose a double skip connection fully connected layer to solve the problems of gradient vanishing and explosion during the training process for the Transformer model, thus further enhancing inference accuracy. The proposed model in this paper achieved an EPE (Absolute error) of 0.57 and 0.61, and a 3PE (Percentage error greater than 3 px) of 1.74% and 1.56% on KITTI 2012 and KITTI 2015 datasets, respectively, with an inference time of 0.46 s and parameters as low as only 2.6 M, showing great advantages compared with other algorithms in various evaluation metrics. [ABSTRACT FROM AUTHOR]

Recent years have witnessed a big leap in automatic visual saliency detection attributed to advances in deep learning, especially Convolutional Neural Networks (CNNs). However, inferring the saliency of each image part separately, as was adopted by most CNNs methods, inevitably leads to an incomplete segmentation of the salient object. In this paper, we describe how to use the property of part-object relations endowed by the Capsule Network (CapsNet) to solve the problems that fundamentally hinge on relational inference for visual saliency detection. Concretely, we put in place a two-stream strategy, termed Two-Stream Part-Object RelaTional Network (TSPORTNet), to implement CapsNet, aiming to reduce both the network complexity and the possible redundancy during capsule routing. Additionally, taking into account the correlations of capsule types from the preceding training images, a correlation-aware capsule routing algorithm is developed for more accurate capsule assignments at the training stage, which also speeds up the training dramatically. By exploring part-object relationships, TSPORTNet produces a capsule wholeness map, which in turn aids multi-level features in generating the final saliency map. Experimental results on five widely-used benchmarks show that our framework consistently achieves state-of-the-art performance. The code can be found on https://github.com/liuyi1989/TSPORTNet. [ABSTRACT FROM AUTHOR]

Deep convolutional neural network is widely used in the field of medical image segmentation. In recent years, multi-scale fusion structure is used to improve segmentation network, which often increases the complexity of the model and reduces the training efficiency while improving the accuracy. To solve these problems, this paper proposed a novel segmentation algorithm for WU-Net pulmonary nodules. It improved the U-Net by introducing an improved residual connection module into the original down-sampling coding channel and mean while using the information channel improved by the new dep module to complete feature extraction and feature fusion. In the experiment, it used LUNA16 dataset to train and verify the model. Meanwhile, comparative experiments used some newly segmentation models. In the experiment with scale of nodules, the Dice coefficient and IoU of proposed model can reach 96.72% and 91.78% respectively.F1-score can reach 92.41% after adding 10% negative samples, which is 1.23% higher than UNet3+.In the experiment with scale of lung parenchyma, the Dice coefficient and IoU of proposed model can reach 83.33% and 66.79% respectively, which is 1.35% and 2.53% higher than RU-Net. The training efficiency of WU-Net is the highest, which is 39.6% higher than U-Net spend. The result shows that WU-Net has improved segmentation effects and training efficiency of model as well. [ABSTRACT FROM AUTHOR]

In order to solve the problems of insufficient feature extraction of sEMG end-to-end gesture recognition and low accuracy of multi-gesture recognition, this paper proposed a multi-stream convolutional EMG gesture recognition network model fusing attention mechanism. The proposed model generated sub-myoelectrograms from multi-channel time-domain sEMG through sliding windows and used a multi-stream convolutional neural network to fully extract the semantic features of each sampled channel sEMG, and then aggregated them to obtain rich multi-channel gesture semantic features. At the same time, the proposed model calculated the attention distribution map of semantic features from temporal and feature channel dimensions to strengthen the useful features and weaken the useless features, further improved the accuracy of multi-gesture recognition. The experiment used the Ninapro dataset for training and testing and compared with the mainstream electromyogram recognition model. The experimental results show that the proposed model has better performance in recognition accuracy, which proves the effectiveness of the proposed model. [ABSTRACT FROM AUTHOR]

The size of the receptive field of the convolutional neural network is related to the size of the convolution kernel. And the traditional convolution uses a fixed-size convolution kernel, which limits the feature perception ability of the network model. In addition, due to the parameter sharing mechanism of the convolutional neural network, it used the same feature extraction method for all pixels in the spatial region. However, there are differences in the distribution of noise signals and clean speech signals in the noisy spectrogram, especially in the complex noise environment, the general convolution method is difficult to achieve high-quality speech signal feature extraction and choosing. In order to solve the above problems, this paper proposed a multi-scale region adaptive convolution module, which used multi-scale information to improve the feature perception ability of the model and automatically allocated the area adaptive convolution achieve and improved the denoising ability of the model. The experiments on the TIMIT public datasets show that the proposed algorithm has achieved satisfactory results in the metrics of speech quality and intelligibility. [ABSTRACT FROM AUTHOR]

During recent years, deep convolutional neural networks (CNNs) have become increasingly popular in sensor-based human activity recognition (HAR), due to powerful feature extraction capacity. However, the required high computational cost has hindered their practical deployment in real-world applications. To address this issue, this paper presents a novel approach named ABCSearch to prune channels through an artificial bee colony (ABC) algorithm, which is a meta-heuristic optimization algorithm inspired by the foraging behavior of honeybees. Unlike traditional pruning strategies that prioritize channel importance, the proposed method seeks to identify the optimal number of channels at each layer, i.e., the optimal pruning structure. To lessen human interference, we transform the search problem of the optimal pruning structure into a simple optimization problem and solve it automatically through the ABC algorithm, which can significantly decrease both model size and computational cost without compromising performance. Specifically, we validate the effectiveness of ABCSearch for a standard five-layer CNN backbone architecture, which achieves 97.05%, 98.45%, 91.71%, 78.89% classification accuracies on four HAR benchmarks, i.e., UCI-HAR, WISDM, PAMAP2, UniMiB-SHAR. Our approach can yield 87.44%, 89.41%, 89.06%, 89.59% FLOPs reductions, and 89.54%, 88.70%, 88.38%, 90.45% less memory consumptions, as well as 4.4 × , 5.0 × , 5.8 × , 6.1 × practical speedups on real hardware devices, while maintaining comparable accuracy. [ABSTRACT FROM AUTHOR]

In recent years, jellyfish outbreaks have frequently occurred in offshore areas worldwide, posing a significant threat to the marine fishery, tourism, coastal industry, and personal safety. Effective monitoring of jellyfish is a vital method to solve the above problems. However, the optical detection method for jellyfish is still in the primary stage. Therefore, this paper studies a jellyfish detection method based on convolution neural network theory and digital image processing technology. This paper studies the underwater image preprocessing algorithm because the quality of underwater images directly affects the detection results. The results show that the image quality is better after applying the three algorithms namely prior defogging, adaptive histogram equalization, and multi-scale retinal enhancement, which is more conducive to detection. We establish a data set containing seven species of jellyfishes and fish. A total of 2141 images are included in the data set. The YOLOv3 algorithm is used to detect jellyfish, and its feature extraction network Darknet53 is optimized to ensure it is conducted in real-time. In addition, we introduce label smoothing and cosine annealing learning rate methods during the training process. The experimental results show that the improved algorithms improve the detection accuracy of jellyfish on the premise of ensuring the detection speed. This paper lays a foundation for the construction of an underwater jellyfish optical imaging real-time monitoring system. [ABSTRACT FROM AUTHOR]

Fault detection is a fundamental and important research topic in automatic seismic interpretation since the geometry of faults usually reveals the accumulation and migration of geological resource. In the communities of machine learning, many methods convert the fault detection task into a patch-based binary classification problem to solve. Because there are only a few faults in seismic data, the related classification task is obviously imbalanced, i.e., non-fault samples heavily outnumber fault samples. On considering that the existing CNN models just transform the problem into a balanced two-class task via undersampling, in this paper we attempt to construct CNNs directly from the imbalanced patch data with the assistance of transfer learning. In particular, we use focal loss to robustly learn fault feature representations for both fault and non-fault patches. With the aim to make full use of available data, we adopt transfer learning to enhance the recognition ability of relatively complex patches. The experiments conducted with both synthetic and real seismic data (some slices of the Netherland offshore F3 block in the North Sea) show that the novel method performs very well in detecting faults in seismic data. In supplemental material, the features extracted by the CNNs are visualized for both fault and non-fault patches to get some insights for their good performance. • We apply CNN with the focal loss to detect seismic faults with imbalanced patches. • Transfer learning strategy is used to fine-tune the CNN to classify complex patches. • We consider more types of faults in generated data to improve the CNN' robustness. • The CNN' predicted results are post-processed by some methods to extract fault lines. [ABSTRACT FROM AUTHOR]

As an essential subproblem in the face recognition field, occluded face recognition has received considerable attention in recent years. However, satisfactory recognition accuracy for occluded faces has not been achieved. Therefore, in order to solve this problem, this paper proposes a convolutional neural network that has multidimensional serial feature extraction modules for occluded faces and uses the deep learning method to improve the recognition rate. In order to improve the expression ability of the network, the method extracts features from two dimensions — space and channel — and built a multidimensional serial feature extraction module. By "multiscale" and "dependence" processing features, the generalization ability was significantly improved. Through the experimental test, the recognition accuracy of the multidimensional feature network (MFNet) reached 90.35%. This accuracy is 17.89% (72.36%) better than that of the traditional algorithm. Meanwhile, compared with other convolutional neural networks, this network has different degrees of improvement, which are 0.68% (ArcFace), 2.14% (Visual Geometry Group Face), and 4.83% (DeepID3). [ABSTRACT FROM AUTHOR]

Semantic segmentation has achieved great success with the popularity of convolutional neural networks (CNNs). However, the huge computational burden restricts the application of most existing networks on edge devices with strict inference time constraints. To solve this problem, a weighted factorized-depthwise convolution network (WFDCNet) is presented in this paper, which contains full- dimensional continuous separation convolution (FCS) modules and a lateral asymmetric pyramid fusion (LAPF) module, aiming to obtain high accuracy without damaging inference speed. Specifically, the FCS module enables the calculation of each dimension to be completed independently in a continuous separation process and uses simplified SE (SSE) attention layer to adjust the channels, achieving the extensive extraction of feature information. The LAPF module is able to eliminate semantic divergence and fuse feature maps of three different scales to realize the combination of multiple information from the front-end and the back-end network. WFDCNet shows superior performance on Cityscapes, Camvid, Mapillary Vistas and COCO-Stuff datasets. Especially, the experimental results demonstrate that our network achieves 73.7% mIoU on Cityscapes dataset, with the inference speed of 102.6FPS on a single RTX 2080 Ti GPU, and 17.2FPS on Jetson TX2. • Separate calculation in all dimensions and use simplified attention layer. • Present the full-dimensional continuous separation convolution module. • Propose the lateral asymmetric pyramid fusion module. • Design an asymmetric encoder-decoder network for real-time semantic segmentation. [ABSTRACT FROM AUTHOR]