Traditional Convolutional Network Research Articles

Career Recommendation for College Students Based on Deep Learning and Machine Learning

With the popularization of higher education, China’s higher education has moved from the stage of elite education to the stage of universal education, and the number of graduates is also increasing. At present, college students are facing tremendous pressure on employment. One is the huge number of jobs, the other is the difference in professional needs, and the third is the spread of job information, which makes it difficult for college students to find a job that suits them. In order to solve this dilemma, this paper analyzes the related technologies and in-depth basic theories of data mining. After introducing several traditional recommendation algorithms, the traditional convolutional network method is improved from three aspects: activation function, pooling strategy, and loss function. Finally, using the hybrid convolutional neural network, a career recommendation model for college students based on deep learning and machine learning is proposed, and simulation experiments are carried out on it. The main research work is as follows: (1) a hybrid convolutional neural network is proposed, which uses convolution operation to learn high-level features to achieve personalized employment recommendation; (2) the training optimization strategy of the hybrid convolutional neural network is studied, aiming at the activation function, pooling processing, and loss function, and the feasibility of the optimization method is verified through simulation experiments; (3) finally, according to the evaluation index of the recommendation algorithm (recall rate and F1-Score), the recall rate of the algorithm in this paper is nearly 15% higher than that of the DNN model. The experiment is compared with the traditional commonly used recommendation algorithm, and the comparative analysis of the experimental results proves the effectiveness of the algorithm for the employment recommendation of college students.

Analysis of Behavioral Image Recognition of Pan-Entertainment of Contemporary College Students’ Network

With the continuous update and iteration of network technology and technological innovation, the handheld smart media of college students will become more and more sensitive. With the advancement of economic globalization, various ideologies and cultures in the world will rapidly invade, and the “pan-entertainment” of online media may intensify. Only through the government’s supervision function and the self-discipline of the internet industry, we can strictly control and screen positive values. In order to better establish the correct employment value orientation of university students and further analyze the importance of the “pan-entertainment” behavior image recognition of college students, this study analyzes the related technology and basic theory of behavior recognition. After introducing several mainstream methods, the traditional dual-stream convolutional network method is improved, and the time information and spatial information extracted by the two channels are discussed for the weighted fusion of feature maps. Finally, using R(2 + 1)D structure and dual-stream network structure design, a deep learning-based spatiotemporal convolution behavior recognition algorithm is proposed. The proposed algorithm is tested and analyzed on the datasets UCF101 and HMDB51. The specific work content is as follows: (1) to summarize the widely used video behavior classification methods proposed so far and discuss the future development. Then, it mainly analyzes the existing technical bottlenecks of some methods based on deep learning methods and summarizes and explores an efficient, stable, and accurate spatiotemporal feature joint extraction and learning method theory. (2) The design of spatiotemporal convolutional network algorithm framework is proposed, the method of segmentation processing of long video is studied, the improvement of the dual-stream network decision-level fusion method is studied, and the R(2 + 1)D network is reorganized. The network algorithm is trained and tested on the UCF-101 dataset and HMDB-51 dataset under the condition of calling the pretrained model. Finally, the accuracy is compared with the existing classic algorithms to obtain better accuracy, which proves the effectiveness of the algorithm for the “pan-entertainment” behavioral image recognition of contemporary college students.

Using the Residual Network Module to Correct the Sub-Seasonal High Temperature Forecast

The high temperature forecast of the sub-season is a severe challenge. Currently, the residual structure has achieved good results in the field of computer vision attributed to the excellent feature extraction ability. However, it has not been introduced in the domain of sub-seasonal forecasting. Here, we develop multi-module daily deterministic and probabilistic forecast models by the residual structure and finally establish a complete set of sub-seasonal high temperature forecasting system in the eastern part of China. The experimental results indicate that our method is effective and outperforms the European hindcast results in all aspects: absolute error, anomaly correlation coefficient, and other indicators are optimized by 8–50%, and the equitable threat score is improved by up to 400%. We conclude that the residual network has a sharper insight into the high temperature in sub-seasonal high temperature forecasting compared to traditional methods and convolutional networks, thus enabling more effective early warnings of extreme high temperature weather.

Breast Cancer Detection and Classification Using Deep CNN Techniques

Breast cancer is a commonly diagnosed disease in women. Early detection, a personalized treatment approach, and better understanding are necessary for cancer patients to survive. In this work, a deep learning network and traditional convolution network were both employed with the Digital Database for Screening Mammography (DDSM) dataset. Breast cancer images were subjected to background removal followed by Wiener filtering and a contrast limited histogram equalization (CLAHE) filter for image restoration. Wavelet packet decomposition (WPD) using the Daubechies wavelet level 3 (db3) was employed to improve the smoothness of the images. For breast cancer recognition, these preprocessed images were first fed to deep convolution neural networks, namely GoogleNet and AlexNet for Adam. Root mean square propagation (RMSprop) and stochastic gradient descent with momentum (SGDM) optimizers were used for different learning rates, such as 0.01, 0.001, and 0.0001. As medical imaging necessitates the presence of discriminative features for classification, the pretrained GoogleNet architectures extract the complicated features from the image and increase the recognition rate. In the latter part of this study, particle swarm optimization-based multi-layer perceptron (PSO-MLP) and ant colony optimization-based multi-layer perceptron (ACO-MLP) were employed for breast cancer recognition using statistical features, such as skewness, kurtosis, variance, entropy, contrast, correlation, energy, homogeneity, and mean, which were extracted from the preprocessed image. The performance of GoogleNet was compared with AlexNet, PSO-MLP, and ACO-MLP in terms of accuracy, loss rate, and runtime and was found to achieve an accuracy of 99% with a lower loss rate of 0.1547 and the lowest run time of 4.14 minutes.

Usis: Unsupervised Semantic Image Synthesis

Semantic Image Synthesis (SIS) is a subclass of image-to-image translation where a photorealistic image is synthesized from a segmentation mask. SIS has mostly been addressed as a supervised problem. However, state-of-the-art methods depend on a huge amount of labeled data and cannot be applied in an unpaired setting. On the other hand, generic unpaired image-to-image translation frameworks underperform in comparison, because they color-code semantic layouts and feed them to traditional convolutional networks, which then learn correspondences in appearance instead of semantic content. In this initial work, we propose a new Unsupervised paradigm for Semantic Image Synthesis (USIS) as a first step towards closing the performance gap between paired and unpaired settings. Notably, the framework deploys a SPADE generator that learns to output images with visually separable semantic classes using a self-supervised segmentation loss. Furthermore, in order to match the color and texture distribution of real images without losing high-frequency information, we propose to use whole image wavelet-based discrimination. We test our methodology on 3 challenging datasets and demonstrate its ability to generate multimodal photorealistic images with an improved quality in the unpaired setting.

Remote Sensing Data Detection Based on Multiscale Fusion and Attention Mechanism

Remote sensing images are often of low quality due to the limitations of the equipment, resulting in poor image accuracy, and it is extremely difficult to identify the target object when it is blurred or small. The main challenge is that objects in sensing images have very few pixels. Traditional convolutional networks are complicated to extract enough information through local convolution and are easily disturbed by noise points, so they are usually not ideal for classifying and diagnosing small targets. The current solution is to process the feature map information at multiple scales, but this method does not consider the supplementary effect of the context information of the feature map on the semantics. In this work, in order to enable CNNs to make full use of context information and improve its representation ability, we propose a residual attention function fusion method, which improves the representation ability of feature maps by fusing contextual feature map information of different scales, and then propose a spatial attention mechanism for global pixel point convolution response. This method compresses global pixels through convolution, weights the original feature map pixels, reduces noise interference, and improves the network’s ability to grasp global critical pixel information. In experiments, the remote sensing ship image recognition experiments on remote sensing image data sets show that the network structure can improve the performance of small-target detection. The results on cifar10 and cifar100 prove that the attention mechanism is universal and practical.

Interpreting Uncertainty in Model Predictions for Covid-19 Diagnosis.

COVID-19, due to its accelerated spread has brought in the need to use assistive tools for faster diagnosis in addition to typical lab swab testing. Chest X-Rays for COVID cases tend to show changes in the lungs such as ground glass opacities and peripheral consolidations which can be detected by deep neural networks. However, traditional convolutional networks use point estimate for predictions, lacking in capture of uncertainty, which makes them less reliable for adoption. There have been several works so far in predicting COVID positive cases with chest X-Rays. However, not much has been explored on quantifying the uncertainty of these predictions, interpreting uncertainty, and decomposing this to model or data uncertainty. To address these needs, we develop a visualization framework to address interpretability of uncertainty and its components, with uncertainty in predictions computed with a Bayesian Convolutional Neural Network. This framework aims to understand the contribution of individual features in the Chest-X-Ray images to predictive uncertainty. Providing this as an assistive tool can help the radiologist understand why the model came up with a prediction and whether the regions of interest captured by the model for the specific prediction are of significance in diagnosis. We demonstrate the usefulness of the tool in chest x-ray interpretation through several test cases from a benchmark dataset.

Fusion of Attentional and Traditional Convolutional Networks for Facial Expression Recognition

INTRODUCTION: The facial expression classification problem has been performed by many researchers. However, it is still a difficult problem to effectively classify facial expressions in highly challenging datasets. In recent years, the self-weighted Squeeze-and-Excitation block (SE-block) technique has evaluated the importance of each feature map in the Convolutional Neural Networks (CNNs) model, corresponding to the output of the Convolution layer, that has shown high efficiency in many practical applications.OBJECTIVES: In this paper, with the aim of balancing speed and accuracy for the problem of facial expression classification, we proposed two novel model architectures to solve these problems.METHODS: Two models proposed in this paper is: (1) a SqueezeNet model combined with a Squeeze-and- Excitation block, (2) SqueezeNet with Complex Bypass combined with a Squeeze-and-Excitation block. These models will have experimented with complex facial expression datasets. Furthermore, the ensemble learning method has also been evidenced to be effective in combining models. Therefore, in order to improve the efficiency of facial expression classification, and aim to compare with the state-of-the-art methods, we use more of the Inception-Resnet V1 model (3). Next, we combine three models (1),(2), and (3) for the classification of facial expressions.RESULTS: The proposed model gives out high accuracy for datasets: namely, with The Extended Cohn-Kanade (CK+) dataset, there are seven basic types of emotions, reaching 99.10 % (using the last 3 frames), 94.20% for the Oulu-CASIA dataset (from 7th frame) with six basic types of emotions, 74.89% for FER2013.CONCLUSION: Experimental results on highly challenging data sets (The Extended Cohn-Kanade, FER2013, Oulu-CASIA) show the effectiveness of the technique of combining three models and two proposed models.

Adaptive Convolution for Semantic Role Labeling

Semantic role labeling (SRL) aims at elaborating the meaning of a sentence by forming a predicate-argument structure. Recent researches depicted that the effective use of syntax can improve SRL performance. However, syntax is a complicated linguistic clue and is hard to be effectively applied in a downstream task like SRL. This work effectively encodes syntax using adaptive convolution which endows strong flexibility to existing convolutional networks. The existing CNNs may help in encoding a complicated structure like syntax for SRL, but it still has shortcomings. Contrary to traditional convolutional networks that use same filters for different inputs, adaptive convolution uses adaptively generated filters conditioned on syntactically-informed inputs. We achieve this with the integration of a filter generation network which generates the input specific filters. This helps the model to focus on important syntactic features present inside the input, thus enlarging the gap between syntax-aware and syntax-agnostic SRL systems. We further study a hashing technique to compress the size of the filter generation network for SRL in terms of trainable parameters. Experiments on CoNLL-2009 dataset confirm that the proposed model substantially outperforms most previous SRL systems for both English and Chinese languages.

Facial Expression Recognition Based on Convolutional Neural Network

At present, facial expression recognition technology is widely used in artificial intelligence, transportation, medical and other aspects, so it has important research value. Traditional facial expression recognition uses manual feature extraction method with low accuracy and weak generalization ability, which is difficult to be applied in real life. With the development of deep learning, convolution neural network appears in people’s vision. Different from traditional manual feature extraction, convolution neural network can learn image features independently, and learn more features. In addition, it has the advantage of sharing the weight with the neural network. Although convolution neural network has multiple advantages, it also has some disadvantages, especially over fitting. In this paper, the model of convolution network is improved based on the classical VGGNet according to the working principle of convolution neural network. In this new model, the number of convolution kernels is reduced in parameters, and the global average pool layer is used to replace the full connection layer in the structure, so as to reduce the degree of over fitting and decrease the operation parameters. Finally, experiments show that the accuracy, generalization and consumption of resource are enhanced in the new model. It is proposed that the new method is better than the traditional convolution network VGGNet.

Adaptive Inception Based on Transfer Learning for Effective Visual Recognition

In computer vision, domain adaptation or transfer learning plays an important role because it learns a target classifier characteristics using labeled data from various distribution. The existing researches mostly focused on minimizing the time complexity of neural networks and it effectively worked on low-level features. However, the existing method failed to concentrate on data augmentation time and cost of labeled data. Moreover, machine learning techniques face difficulty to obtain the large amount of distributed label data. In this research study, the pre-trained network called inception layer is fine-tuned with the augmented data. There are two phases present in this study, where the effectiveness of data augmentation for Inception pre-trained networks is investigated in the first phase. The transfer learning approach is used to enhance the results of the first phase and the Support Vector Machine (SVM) is used to learn all the features extracted from inception layers. The experiments are conducted on a publicly available dataset to estimate the effectiveness of proposed method. The results stated that the proposed method achieved 95.23% accuracy, where the existing techniques namely deep neural network and traditional convolutional networks achieved 87.32% and 91.32% accuracy respectively. This validation results proved that the developed method nearly achieved 4-8% improvement in accuracy than existing techniques.

A novel approach with dual-sampling convolutional neural network for ultrasound image classification of breast tumors

Breast cancer is one of the leading causes of female cancer deaths. Early diagnosis with prophylactic may improve the patients’ prognosis. So far ultrasound (US) imaging has been a popular method in breast cancer diagnosis. However, its accuracy is bounded to traditional handcrafted feature methods and expertise. A novel method, named dual-sampling convolutional neural networks (DSCNNs), was proposed in this paper for the differential diagnosis of breast tumors based on US images. Combining traditional convolutional and residual networks, DSCNN prevented gradient disappearance and degradation. The prediction accuracy was increased by the parallel dual-sampling structure, which can effectively extract potential features from US images. Compared with other advanced deep learning methods and traditional handcrafted feature methods, DSCNN reached the best performance with an accuracy of 91.67% and an area under curve of 0.939. The robustness of the proposed method was also verified by using a public dataset. Moreover, DSCNN was compared with evaluation from three radiologists utilizing US-BI-RADS lexicon categories for overall breast tumors assessment. The result demonstrated that the prediction sensitivity, specificity and accuracy of the DSCNN were higher than those of the radiologist with 10 year experience, suggesting that the DSCNN has the potential to help doctors make judgements in clinic.

Identification of Microseismic Events in Rock Engineering by a Convolutional Neural Network Combined with an Attention Mechanism

Microseismic technology has widely been used in many rock engineering applications to shield workers from engineering hazards and monitor underground construction. To avoid the heavy workloads imposed by the manual recognition of many microseismic signals, this study proposes a new end-to-end training network architecture to automatically identify microseismic events. A dataset including not only easily identifiable microseismic signals but also barely distinguishable nontypical data has been collected from a practical rock engineering project for training and testing the network model. The applicability of various networks for this task is discussed to select the best method for microseismic recognition. We modify the residual skip connections to make them more suitable for the signal classification task. Then, the novel depthwise spatial and channel attention (DSCA) module is proposed. This module can autonomously learn how to weight information with different levels of importance, similar to human attention, which greatly improves the network performance without incurring additional computational costs. Theoretically, it can be a useful tool to replace traditional denoising algorithms and model the interdependencies between the different channels of a multichannel signal. Furthermore, the DSCA module and the modified residual connections are combined with a traditional convolutional network to obtain a novel network architecture named ResSCA and the results of comparative experiments are presented. Finally, single- and multichannel models are constructed based on ResSCA, which achieved improved accuracy rates. Their advantages and drawbacks are analyzed. This study presents a modified network architecture suitable for identifying and classifying complex signals to enable intelligent microseismic monitoring, which is valuable for various rock engineering applications.

Lung diseases identification method based on capsule neural network

In the automatic identification of lesions, aiming at the misdiagnosis of the lung nodule size, shape, blood vessels and other lung tissues on CT images of pneumonia, this paper proposes a method for identifying lung diseases based on capsule neural networks. Considering that the texture features of lung CT images contain important medical information, this article first combines gray average, entropy, fractal dimension and fractal intercept to form feature vectors as texture features, and introduces context models to obtain context information. The LBG algorithm was used to achieve context quantization, and the CT images of lungs were preprocessed. Then, for the problem of less CT image data of lungs and fewer features of lung diseases in this paper, the existing images are enhanced and the data set is expanded. Finally, in the recognition process of pneumonia, the characteristic pose information will affect the recognition. In order to solve this problem, this paper uses a capsule neural network to identify and detect lung CT images. The application results show that the accuracy rate of identifying lung diseases by capsule neural network is 98.7%. Compared with traditional convolutional networks, capsule neural networks can better identify lung diseases.

A Convolutional Neural Network with Parallel Multi-Scale Spatial Pooling to Detect Temporal Changes in SAR Images

In synthetic aperture radar (SAR) image change detection, it is quite challenging to exploit the changing information from the noisy difference image subject to the speckle. In this paper, we propose a multi-scale spatial pooling (MSSP) network to exploit the changed information from the noisy difference image. Being different from the traditional convolutional network with only mono-scale pooling kernels, in the proposed method, multi-scale pooling kernels are equipped in a convolutional network to exploit the spatial context information on changed regions from the difference image. Furthermore, to verify the generalization of the proposed method, we apply our proposed method to the cross-dataset bitemporal SAR image change detection, where the MSSP network (MSSP-Net) is trained on a dataset and then applied to an unknown testing dataset. We compare the proposed method with other state-of-arts and the comparisons are performed on four challenging datasets of bitemporal SAR images. Experimental results demonstrate that our proposed method obtain comparable results with S-PCA-Net on YR-A and YR-B dataset and outperforms other state-of-art methods, especially on the Sendai-A and Sendai-B datasets with more complex scenes. More important, MSSP-Net is more efficient than S-PCA-Net and convolutional neural networks (CNN) with less executing time in both training and testing phases.

Recognition of Fusion Processing Based on Infrared and Visible Images under the Framework of Semi-supervised Learning

In traditional convolutional networks, due to the lack of adequate information protection of traditional image fusion technology and the incomplete removal of redundant noise, the useful information of the fusion image is missing and the recognition success rate is low. In this paper, through the research of deep learning-based image fusion methods and traditional target recognition and SVM neural network, an image fusion processing recognition method based on infrared and visible light is designed. The coding network of this image fusion method consists of convolutional layers, fusion layers and dense blocks. The output of each layer needs to be connected to the next few layers by using a densely connected neural network, so as to obtain more useful features from the source image and fuse the data of the two images better. It is verified by simulation that the fused image has sound visual effects, and its edges and details have been completely preserved. Thus, the target object has strong recognizability compared with the surrounding environment. Research shows that this method will help to more accurately interpret target information in complex environments and achieve more effective results in target recognition.

Learning visual features under motion invariance

Humans are continuously exposed to a stream of visual data with a natural temporal structure. However, most successful computer vision algorithms work at image level, completely discarding the precious information carried by motion. In this paper, we claim that processing visual streams naturally leads to formulate the motion invariance principle, which enables the construction of a new theory of learning that originates from variational principles, just like in physics. Such principled approach is well suited for a discussion on a number of interesting questions that arise in vision, and it offers a well-posed computational scheme for the discovery of convolutional filters over the retina. Differently from traditional convolutional networks, which need massive supervision, the proposed theory offers a truly new scenario for the unsupervised processing of video signals, where features are extracted in a multi-layer architecture with motion invariance. While the theory enables the implementation of novel computer vision systems, it also sheds light on the role of information-based principles to drive possible biological solutions.

Small Object Detection with Multiple Receptive Fields

Small object detection has been a problem in deep learning convolutional neural network models. A multi-rate dilated convolution module is proposed to form a feature map to locate small objects. Benefiting from the fact that the dilated convolution does not add extra complexity while maintaining the characteristics of the high-resolution feature map, this paper replaces the traditional convolution network by setting the dilated convolution with the ratio of 1, 2, and 5, and fuses the different convolution rate convolutions. The feature map branch, extracting high-level features, using the FPN algorithm as the basic network framework, combining a variety of receptive field feature maps, enhancing the detection ability of the algorithm for small objects, and the convergence accuracy is greatly improved. Experiments show that the proposed method achieves 81.9% MAP in the PASCAL VOC2007 dataset, which exceeds the traditional object detection algorithm.

Efficiency Optimization of Capsule Network Model Based on Vector Element

Currently, Deep Learning and Convolutional Neural Network (CNN) have been widely used in many fields and have generated very high value in these fields, especially in the field of image recognition. But there are some deficiencies in certain issues of image recognition. For example, CNN’s recognizing performance is not good at different angles of objects and overlapping objects. Also, CNN is sometimes very sensitive to slight perturbations, modifying one pixel of a recognized image may cause recognition errors. For these problems, the capsule network (CapsNet) proposed by Geoffrey Hinton can solve the problems of traditional convolutional networks. Shortly after CapsNet was proposed, the model structure was relatively simple, and many aspects could be explored for improvement. This paper will optimize CapsNet from two aspects: “optimization of routing mechanism” and “increase Dropout operation.” And carry out experiments and results analysis on these optimizations.

Study on MRI Medical Image Segmentation Technology Based on CNN-CRF Model

Image segmentation is an important technique for segmenting images without overlapping each other and having their own features. It has been rapidly developed in the field of medical imaging, but there is currently a difference between classification accuracy and segmentation accuracy for medical image segmentation. In this paper, the deep convolutional neural network is combined with the cascading structure, and a uniform learning framework is established with the use conditional random field. This paper first adds a cascading structure under the deep convolutional neural networks (DCNN) framework to more effectively simulate the direct dependencies between spatial closure tags. Secondly, the conditional random field (CRF) is used for post-segmentation processing, which effectively solves the contradiction between the segmentation accuracy and the network depth and the number of pooling times in the traditional convolutional network. Secondly, the CRF is used for post-segmentation processing, which effectively solves the contradiction between the segmentation accuracy and the network depth and the number of pooling times in the traditional convolutional network.