Search

There are several unresolved issues in the field of ship instance segmentation in synthetic aperture radar (SAR) images. First, in inshore dense ship area, the problems of missed detections and mask overlap frequently occur. Second, in inshore scenes, false alarms occur due to strong clutter interference. In order to address these issues, we propose a novel ship instance segmentation network based on dynamic key points information enhancement. In the detection branch of the network, a dynamic key points module is designed to incorporate the target's geometric information into the parameters of the dynamic mask head using an implicit encoding technique. In addition, we introduce a dynamic key points encoding branch, which encodes the target's strong scattering regions as dynamic key points. It strengthens the network's ability to learn the correspondence between local regions with strong scattering and overall ship targets, effectively mitigating mask overlap issues. Moreover, it enhances the discriminative ability of network between ship targets and clutter interference, leading to a reduction in false alarm rates. To further enhance the dynamic key points information, an instancewise attention map module is designed, which decodes the key points during the mask prediction period, generating instancewise attention maps based on 2-D Gaussian distribution. This module further enhances the sensibility of network to specific instances. Simulation experiments conducted on the Polygon Segmentation SAR Ship Detection Dataset and High-Resolution SAR Images Dataset demonstrate the superiority of our proposed method over other state-of-the-art methods in inshore and offshore scenes.

In order to realize waypoint-based automatic berthing alongside or astern, a generation algorithm of dynamic virtual ships is proposed to guide the underactuated surface vehicle (USV). In this paper, a nonlinear model predictive control (NMPC) problem is set up with multiple shooting method, which lifts the system and improves convergency. Measuring error and disturbances are reduced by moving horizon estimation (MHE) where the weights of sensors are auto-tuned according to their variation. Simulations of berthing alongside and astern of ship BAY CLASS illustrates the position error is smaller than 0.6 m and course error is smaller than 0.5°. Compared to adaptive sliding mode control, proposed method has advantage on state smoothness and the control performance meets the requirements of berthing accuracy.

In this paper, a new method is proposed and verified to eliminate the non-ideal resonance caused by DGS application in GCPW by adding via holes to reduce electromagnetic leakage. By adding a dense through-hole array around the dumbbell DGS can form an electrical wall to confine the electromagnetic to the central conductor. The test results show that the non-ideal resonance is eliminated after adding the through hole and a wide stopband from 16.1 GHz to 17.8 GHz with rejection larger than 20 dB. A double-dumbbell bandpass filter was successfully designed with this tech eliminating the non-ideal resonance at 8.053 GHz and 14.232 GHz. The results show that this structure can effectively eliminate the irrational resonance caused by electromagnetic leakage.

With the development of computer vision technology, the demand for accurate recognition of stereoscopic image quality in the market is increasing. Accurate recognition of stereoscopic image quality is of great significance for providing high-value intelligent image services, which is also the motivation for this study. Accurately recognizing the quality of stereoscopic images is of great significance for image analysis and computer vision applications. In fields such as autonomous driving, medical image analysis, and industrial detection, accurate stereo image quality can provide reliable input for algorithms and improve the accuracy of analysis and recognition. The field or issue of this study is image quality evaluation, which aims to find higher performance stereoscopic image quality evaluation methods. Therefore, this study draws inspiration from the idea of ensemble learning and designs two Convolutional Neural Network (CNN) stereoscopic image quality evaluation methods based on the semantic features of stereoscopic images and the local detail perception module, and fuses them to form a mixed evaluation model. This study aims to solve the problem of image quality assessment, which is to accurately identify the quality of stereoscopic images and provide high-value intelligent image services. With the continuous development of computer vision technology, the demand for accurate recognition of stereoscopic image quality is increasing, which is also the motivation of this study. This study drew inspiration from the idea of ensemble learning and designed two hybrid evaluation models based on stereo image semantic features and local detail perception modules. Convolutional Neural Network (CNN) was used to achieve stereo image quality evaluation. This is one of the main contributions of this article. In order to evaluate the performance of the designed model, the LIVE 3D Phase I dataset was used for testing experiments. The expected results show that when the number of test samples is 500, the overall measurement values of the Spearman rank ordered correlation coefficient (SROCC) and Pearson linear correlation coefficient (PLCC) of the designed ICNN1 and ICNN2 stereo image quality evaluation models are 0.940, 0.949, and 0.940, 0.949, respectively. These results are significantly higher than the selected contrastive deep learning models and machine learning models. In addition, the designed model has relatively low computational time but high computational memory consumption, which is one of the main gaps compared to other studies. In summary, the model designed in this study has great application potential in improving the accuracy of stereo image quality recognition, and is particularly suitable for the Chinese visual design industry. Future research can further explore the market-oriented application of this recognition model.

Image steganalysis aims to detect whether secret information is hidden in images and is a means to solve the communication security. Recently, a series of convolutional neural network-based steganalysis models has been proposed and has achieved remarkable performance. Given that the steganalysis task is different from other computer vision tasks, the series of models is inevitably restricted by manual factors, and some other useful residual information will be missed in the process of feature extraction. To solve this problem, this paper proposes a network model based on non-local operations and multi-channel convolution as part of the basic block of feature extraction for spatial grayscale image steganalysis. The preprocessing layer is improved to compress the image content, further using the introduced non-local operations and multi-channel convolution modules to enhance the residual information in high-frequency regions and extract diverse steganographic features effectively. Extensive ablation study shows that the introduced extraction module can improve the steganographic detection accuracy effectively. Our method is generally better than other methods in terms of detection accuracy, especially in low payload detection.

Language-modeling-based methods for story plot generation aim to generate a plot with a language model (LM). LM methods have limitations of user-assist plot generation of goal control, refinement for editing, causing the generated plots not clear sense for specific goal, lack coherence, and edit flexible. We present a control-and-edit transformer technique which uses controlled imitation learning of editing distance from dynamic programming to support deleting policy, inserting policy, a weighting-reward with prepossess of corpus statistic, and measures continues reward for the controlled goal. Automated evaluation and Haman judgement show our method is promising in comparison with the baselines.