Wenyuan Yang

Bio: Wenyuan Yang is an academic researcher from Zhangzhou Normal University. The author has contributed to research in topics: Video tracking & Computer science. The author has an hindex of 3, co-authored 5 publications receiving 19 citations.

Small object detection via dual inspection mechanism for UAV visual images

Abstract: Unmanned Aerial Vehicles (UAVs) are utilized instead of humans to complete aerial assignments in various fields. With the development of computer vision, object detection has become one of the core technologies in UAV application. However, object detection of small targets often has missed detection, and the detection performance is far less than that of large targets. In this paper, we propose a dual inspection mechanism, which identifies missed targets in suspicious areas to assist single-stage detection branches, and shares dual decisions to make feature-level multi-instance detection modules produce reliable results. Firstly, the detection results contain missed targets is confirmed, which are in the part that does not reach the confidence threshold. For this reason, the feature vector provided by the denoising sparse autoencoder is calculated, and this part of the result is filtered again. Secondly, we empirically reveal that single detection results are not reliable enough, and the multiple attributes of the target need to be considered. Motivated by this, the initial and secondary detection results are combined and rank by importance. Finally, we give the corresponding confidence to the top-ranked instance, making it possible to become the object again. Experimental results reflect that our mechanism improves 2.7% mAP on the VisDrone2020 dataset, 1.0% mAP on the UAVDT dataset and 1.8% mAP on the MS COCO dataset. We propose detection mechanism which achieves state-of-the-art levels on these datasets and it performs better on small object detection.

Supervised Learning via Unsupervised Sparse Autoencoder

Abstract: Dimensionality reduction is commonly used to preprocess high-dimensional data, which is an essential step in machine learning and data mining. An outstanding low-dimensional feature can improve the efficiency of subsequent learning tasks. However, existing methods of dimensionality reduction mostly involve datasets with sufficient labels and fail to achieve effective feature vectors for datasets with insufficient labels. In this paper, an unsupervised multiple layered sparse autoencoder model is studied. Its advantage is that it reduces the reconstruction error as its optimization goal, with the resulting low-dimensional feature being reconstructed to the original dataset as much as possible. Therefore, the reduction of high-dimensional datasets to low-dimensional datasets is effective. First, the relationship among the reconstructed data, the number of iterations, and the number of hidden variables is explored. Second, the dimensionality reduction ability of the sparse autoencoder is proven. Several classical feature representation methods are compared with the sparse autoencoder on publicly available datasets, and the corresponding low-dimensional representations are placed into different supervised classifiers and the classification performances reported. Finally, by adjusting the parameters that might influence the classification performance, the parametric sensitivity of the sparse autoencoder is shown. The extensively low-dimensional feature classification experimental results demonstrated that the sparse autoencoder is more efficient and reliable than the other selected classical dimensional reduction algorithms.

UAV object tracking by background cues and aberrances response suppression mechanism

Abstract: Real-time object tracking for unmanned aerial vehicles (UAVs) is an essential and challenging research topic for computer vision. However, the scenarios that UAVs deal with are complicated, and the UAV tracking targets are small. Therefore, general trackers often fail to take full advantage of their performance in UAV scenarios. In this paper, we propose a tracking method for UAV scenes, which utilizes background cues and aberrances response suppression mechanism to track in 4 degrees-of-freedom. Firstly, we consider the tracking task as a similarity measurement problem. In this study, we decompose this problem into two subproblems for optimization. Secondly, to alleviate the problem of small targets in UAV scenes, we utilize background cues fully. Also, to reduce interference by background information, we employ an aberrance response suppression mechanism. Then, to obtain accurate target state information, we introduce a logarithmic polar coordinate system. We perform phase correlation calculations in logarithmic polar coordinates to obtain the rotation and scale changes of the target. Finally, target states are obtained through response fusion, which includes displacement, scale, and rotation angle. Our approach is carried out in a large number of experiments on various UAV datasets, such as UAV123, DBT70, and UAVDT2019. Compared with the current advanced trackers, our method is superior on UAV small target tracking.

Parallel Dual Networks for Visual Object Tracking

Abstract: Visual Object Tracking plays an essential role in solving many basic problems in computer vision. In order to improve the tracking accuracy, the previous methods have prevented tracking failures from occurring by improving the ability to describe the target. However, few of them consider ways to relocate and track the target after a tracking failure. In this paper, we propose the use of a parallel dual network for visual object tracking. This is constructed from two networks and an adjustment module to enable judgement of tracking failures, as well as target relocation and tracking. Firstly, we employ the Siamese matching method and correlation filter method to build tracking network and inspecting network. Both networks track the target simultaneously to obtain two tracking results. Secondly, an adjustment module is constructed, which compares the overlap ratio of the two tracking results with a set threshold, then fuses them or selects the best one. Finally, the fusion or selection result is output and the tracker is updated. We perform comprehensive experiments on five benchmarks: VOT2016, UAV123, Temple Color-128, OTB-100 and OTB-50. The results demonstrate that, compared with other state-of-the-art algorithms, the proposed tracking method improves tracking precision while maintaining real-time performance.

PEMFC Residual Life Prediction Using Sparse Autoencoder-Based Deep Neural Network

Abstract: In the cause of working out the challenge of remaining life prediction (RUL) of proton exchange membrane fuel cell (PEMFC) under dynamic operating conditions, this article proposes a PEMFC RUL forecast technique based on the sparse autoencoder (SAE) and deep neural network (DNN). The method extracts the data set from the original experimental data at intervals periods of one hour to realize datum reconstruction. The Gaussian-weighted moving average filter is used to smooth noisy data (voltage and current). The smoothed filtered power output signal of the stack is extracted as an aging indicator. The SAE is used to extract the prediction features automatically, and the DNN is applied to realize the RUL prediction. The proposed method is experimentally verified using 127 369 experimental data. The effectiveness of the novel method is verified by three different training sets and test set configurations. The experimental results reveal that the novel approach has the best prediction effectiveness when the training set length is set to 500 h. At this point, the prediction accuracy can reach 99.68%. The mean absolute error (MAE), mean square error (MSE), and root-mean-square error (RMSE) are minimum values, which are 0.2035, 0.1121, and 0.3348, respectively. The superiority and effectiveness of the proposed approach are further validated by comparison with the K-nearest neighbor and support vector regression machine. The proposed approach can be appropriate for the prediction of the RUL of PEMFC under dynamic conditions.

Improved Heart Disease Prediction Using Particle Swarm Optimization Based Stacked Sparse Autoencoder

Abstract: Heart disease is the leading cause of death globally. The most common type of heart disease is coronary heart disease, which occurs when there is a build-up of plaque inside the arteries that supply blood to the heart, making blood circulation difficult. The prediction of heart disease is a challenge in clinical machine learning. Early detection of people at risk of the disease is vital in preventing its progression. This paper proposes a deep learning approach to achieve improved prediction of heart disease. An enhanced stacked sparse autoencoder network (SSAE) is developed to achieve efficient feature learning. The network consists of multiple sparse autoencoders and a softmax classifier. Additionally, in deep learning models, the algorithm’s parameters need to be optimized appropriately to obtain efficient performance. Hence, we propose a particle swarm optimization (PSO) based technique to tune the parameters of the stacked sparse autoencoder. The optimization by the PSO improves the feature learning and classification performance of the SSAE. Meanwhile, the multilayer architecture of autoencoders usually leads to internal covariate shift, a problem that affects the generalization ability of the network; hence, batch normalization is introduced to prevent this problem. The experimental results show that the proposed method effectively predicts heart disease by obtaining a classification accuracy of 0.973 and 0.961 on the Framingham and Cleveland heart disease datasets, respectively, thereby outperforming other machine learning methods and similar studies.

Mutual Learning and Feature Fusion Siamese Networks for Visual Object Tracking

Abstract: Recently Siamese-based trackers have shown their outstanding performance in visual object tracking community. But they seldom pay attention to the inter-branch interaction as well as intra-branch feature fusion from different convolution layers. In this paper, we build up a comprehensive Siamese network which consists of a mutual learning subnetwork (M-net) and a feature fusion subnetwork (F-net), to realize object tracking. Each of them is a Siamese network with special functions. M-net is designed to help the two branches mine the dependencies from each other, thus the object template is adaptively updated to a certain extent. F-net fuses different levels of convolutional features for full usage of spatial and semantic information. We also design a global-local channel attention (GLCA) module in F-net to capture the channel dependencies for a proper feature fusion. Our method takes ResNet as feature extractor and is trained offline in an end-to-end style. We evaluate our method in several famous benchmarks such as OTB2013, OTB2015, VOT2015, VOT2016, NFS and TC128. Extensive experimental results demonstrate our method achieves competitive results while maintaining a considerable real-time speed.

Magnetic flux leakage image classification method for pipeline weld based on optimized convolution kernel

Abstract: In order to intelligently classify magnetic flux leakage signals, this study proposes a method of magnetic flux leakage image classification based on sparse self-coding. With inputting with the magnetic flux leakage image of the pipe weld, it extracts features automatically from the Convolutional Neural Network (CNN) rather than the artificial extraction process. The network classification ability can be improved through pre-training of the convolution kernel and introducing the sparse constraints and the image entropy similarity constraint rules. The experiment uses 500 images of magnetic flux leakage signals to classify the girth welds and spiral welds. The accuracy of classification is 95.1%, which is superior to the traditional convolution neural network model. Experimental results show that the improved model has good feature extraction ability and generalization ability.