Deep Neural Network Based Detection and Segmentation of Ships for Maritime Surveillance

In recent years,computer visionfinds wide applications in maritime surveillance with its sophisticated algorithms and advanced architecture.Auto-matic ship detection with computer vision techniques provide an efficient means to monitor as well as track ships in water bodies.Waterways being an important medium of transport require continuous monitoring for protection of national security.The remote sensing satellite images of ships in harbours and water bodies are the image data that aid the neural network models to localize ships and to facilitate early identification of possible threats at sea.This paper proposes a deep learning based model capable enough to classify between ships and no-ships as well as to localize ships in the original images using bounding box tech-nique.Furthermore,classified ships are again segmented with deep learning based auto-encoder model.The proposed model,in terms of classification,provides suc-cessful results generating 99.5%and 99.2%validation and training accuracy respectively.The auto-encoder model also produces 85.1%and 84.2%validation and training accuracies.Moreover the IoU metric of the segmented images is found to be of 0.77 value.The experimental results reveal that the model is accu-rate and can be implemented for automatic ship detection in water bodies consid-ering remote sensing satellite images as input to the computer vision system.

Hformer:highly efficient vision transformer for low-dose CT denoising

In this paper,we propose Hformer,a novel supervised learning model for low-dose computer tomography(LDCT)denoising.Hformer combines the strengths of convolutional neural networks for local feature extraction and transformer models for global feature capture.The performance of Hformer was verified and evaluated based on the AAPM-Mayo Clinic LDCT Grand Challenge Dataset.Compared with the former representative state-of-the-art(SOTA)model designs under different architectures,Hformer achieved optimal metrics without requiring a large number of learning parameters,with metrics of33.4405 PSNR,8.6956 RMSE,and 0.9163 SSIM.The experiments demonstrated designed Hformer is a SOTA model for noise suppression,structure preservation,and lesion detection.

Multiple input self-organizing-map ResNet model for optimization of petroleum refinery conversion units

This work introduces a deep-learning network,i.e.,multi-input self-organizing-map ResNet(MISR),for modeling refining units comprised of two reactors and a separation train.The model is comprised of self-organizing-map and the neural network parts.The self-organizing-map part maps the input data into multiple two-dimensional planes and sends them to the neural network part.In the neural network part,residual blocks enhance the convergence and accuracy,ensuring that the structure will not be overfitted easily.Development of the MISR model of hydrocracking unit also benefits from the utilization of prior knowledge of the importance of the input variables for predicting properties of the products.The results show that the proposed MISR structure predicts more accurately the product yields and properties than the previously introduced self-organizing-map convolutional neural network model,thus leading to more accurate optimization of the hydrocracker operation.Moreover,the MISR model has smoother error convergence than the previous model.Optimal operating conditions have been determined via multi-round-particle-swarm and differential evolution algorithms.Numerical experiments show that the MISR model is suitable for modeling nonlinear conversion units which are often encountered in refining and petrochemical plants.

基于多通道图卷积网络的节点聚类

针对在深度聚类中大部分基于图卷积网络(graph convolutional network,GCN)的方法仅使用拓扑图而忽略了特征空间中存在的结构信息的问题,提出一种通过引入特征图更充分地利用特征空间中存在的结构信息的节点聚类方法.首先,该方法使用自动编码器(auto-encoder,AE)来学习节点特征的潜在表示,同时在特征图、拓扑图及节点属性3个层面获得节点嵌入;然后,使用融合机制对学习到的节点嵌入进行融合;最后,通过自监督的方式训练网络实现节点聚类.在6个基准数据集上的大量实验表明,该方法明显提高了聚类精度.

Fault diagnosis for distillation process based on CNN–DAE

Distillation is the most widely used operation for liquid mixture separation in the chemical industry. It is of great importance to detect and diagnose faults in distillation process. Due to the strong feedback and coupling of processes in a distillation column, it is difficult to use deep auto-encoders(DAEs) alone to achieve good results in detecting and diagnosing faults, in terms of accuracy and efficiency. This paper proposes a hybrid fault-diagnosis model based on convolutional neural networks(CNNs) and DAEs, by integrating the powerful capability of CNN in feature extraction and of DAE in classification. A case study was carried out with the distillation process of depropanization. It is shown that the proposed hybrid model is of good performance compared to other models, in terms of the accuracy of fault detection in such a process. Also, with the increase of structural layers of the CNN–DAE model, the diagnostic accuracy will be improved, with an optimal accuracy of 92.2%.

Image label transfer: Short video labelling by using frame auto-encoder

Short videos on the Internet have a huge amount, but most of them are unlabeled. In this paper, a rough short video labelling method based on the image classification neural network is proposed. Convolutional auto-encoder is applied to train and learn unlabeled video frames, in order to obtain feature in the specific level. With these features, the video key-frames are extracted by the feature clustering method. These key-frames which represent the video content are put into an image classification network, so that the labels of every video clip can be got. In addition, the different architectures of convolutional auto-encoder are estimated, and a better performance architecture through the experiment result is selected. In the final experiment, the video frame features from the convolutional auto-encoder are compared with those from other extraction methods, where it illustrates remarkable results by the proposed method.

A denoising-classification neural network for power transformer protection

Artificial intelligence(AI)can potentially improve the reliability of transformer protection by fusing multiple features.However,owing to the data scarcity of inrush current and internal fault,the existing methods face the problem of poor generalizability.In this paper,a denoising-classification neural network(DCNN)is proposed,one which inte-grates a convolutional auto-encoder(CAE)and a convolutional neural network(CNN),and is used to develop a reli-able transformer protection scheme by identifying the exciting voltage-differential current curve(VICur).In the DCNN,CAE shares its encoder part with the CNN,where the CNN combines the encoder and a classifier.Based on the inter-action of the CAE reconstruction process and the CNN classification process,the CAE regards the saturated features of the VICur as noise and removes them accurately.Consequently,it guides CNN to focus on the unsaturated features of the VICur.The unsaturated part of the VICur approximates an ellipse,and this significantly differentiates between a healthy and faulty transformer.Therefore,the unsaturated features extracted by the CNN help to decrease the data ergodicity requirement of AI and improve the generalizability.Finally,a CNN which is trained well by the DCNN is used to develop a protection scheme.PSCAD simulations and dynamic model experiments verify its superior performance.