Ship object detection faces the challenge of increasing the difficulty of positioning in hazy environments. Additionally, the latest convolutional neural network (CNN) cannot obtain satisfactory detection results. Therefore, we propose ERDNet, a dual-branch-driven end-to-end network, to improve ship detection accuracy during hazy weather. Specifically, we design a two-branch feature extraction network through complementary attentional fusion to enhance the object feature information of low-quality images. Second, we designed a feature pyramid fusion structure called ERPSA-PAN to aggregate context information effectively. ERPSA-PAN improves the feature fusion capability of the model by suppressing background interference and enhancing useful information. In addition, the spatial-frequency fusion block (SFFB) module with expanded receptive fields is added to the ERDNet detection head to improve the detection ability for multiscale targets. More importantly, we design a robust haze loss to handle different degrees of haze. We introduce two new haze ship datasets, Hazy-SeaShips and Hazy-Boats, which include 17,000 synthetic haze images and 2898 real haze images, respectively, to address the lack of hazy ship image datasets. The images cover variations such as haze thicknesses, ship types, and scales, along with complex backgrounds, and occlusions. The experimental results show that the proposed method is superior to other state-of-the-art (SOTA) methods and achieves relatively competitive results. The source codes, and datasets are available on https://github.com/ZikHH/ERDNet.

Video stabilization aims to eliminate random jitter in video sequences, but most methods result in stabilized video with degraded resolution and content loss. In this letter, we propose a full-frame video stabilization algorithm based on motion intent analysis. The algorithm consists of three main steps: motion estimation, motion smoothing, and video completion. First, robust keypoints are extracted using the improved SuperPoint network and refined with the suppression via square covering (SSC) algorithm to obtain stable and reliable keypoints. Then, the Lucas-Kanade algorithm is applied for motion estimation of inter-frame matched feature points. Second, motion smoothing is achieved using the Kalman filtering algorithm to remove the high-frequency jitter component from the trajectory, and motion compensation is applied to the original video sequence to generate a stable image sequence. Finally, to preserve the original video resolution, we propose a video completion method based on motion intent analysis. Experimental results demonstrate that our method achieves higher stability while maintaining the original video resolution compared to the current state-of-the-art video stabilization algorithms.

Segmenting objects from cluttered backgrounds in single-channel images, such as marine radar echoes, medical images, and remote sensing images, poses significant challenges due to limited texture, color information, and diverse target types. This paper proposes a novel solution: the Onet, an O-shaped assembly of twin U-Net deep neural networks, designed for unsupervised binary semantic segmentation. The Onet, trained with an intensity-complementary image pair and without the need for annotated labels, maximizes the Jensen-Shannon divergence (JSD) between the densely localized features and the class probability maps. By leveraging the symmetry of U-Net, Onet subtly strengthens the dependence between dense local features, global features, and class probability maps during the training process. The design of the complementary input pair aligns with the theoretical requirement that optimizing JSD needs the class probability of negative samples to accurately estimate the marginal distribution. Compared to the current leading unsupervised segmentation methods, the Onet demonstrates superior performance in target segmentation in marine radar frames and cloud segmentation in remote sensing images. Notably, we found that Onet's foreground prediction significantly enhances the signal-to-noise ratio (SNR) of targets amidst marine radar clutter. Onet's source code is publicly accessible at https://github.com/joeyee/Onet.

Different from previous literature assuming either single reconfigurable intelligent surface (RIS) or multiple RISs at given locations, this letter first studies a multi-user millimeter wave (mmWave) massive multiple-input multiple-output (MIMO) system aided by multiple RISs following Poisson point processes. Then we derive an accurate closed-form expression for the spectral efficiency of an arbitrary user and an approximation for the ergodic spectral efficiency of the cell. Moreover, we consider the hardware impairment and model a practical phase-dependent amplitude for each reflecting element. Finally, the simulation results verify the derivations and demonstrate that RISs can provide greater performance gain than the active nodes and require different deployment strategy, which is related to the number of RISs.

Ship detection in Synthetic Aperture Radar (SAR) images has been widely applied in the military and civil fields. However, the background environment of SAR images is complex and there are many interferences similar to the ship targets, which is easy to lead fault detection and affect the detection performance. To address this problem, a vision-inspired filtering algorithm (FilterGAN) is proposed to filter out the target-irrelevant information. Firstly, we build a representation model based on filtering mechanism of human brain to guide the design of filtering network. Secondly, to simulate the adjustment process of the priority map reconstruction in human brain, Generative Adversarial Networks (GAN) is used to learn the optimal filtering mapping function. To train FilterGAN, we introduce the labeling process to generate the ground-truth filtered SAR image. Experimental results on AIR-SARShip-1.0 dataset demonstrate that the detection performance of SAR ships can be improved obviously with FilterGAN.