3D object detection in LiDAR point clouds has been extensively used in autonomous driving, intelligent robotics, and augmented reality. Although the one-stage 3D detector has satisfactory training and inference speed, there are still some performance problems due to insufficient utilization of bird’s eye view (BEV) information. In this paper, a new backbone network is proposed to complete the cross-layer fusion of multi-scale BEV feature maps, which makes full use of various information for detection. Specifically, our proposed backbone network can be divided into a coarse branch and a fine branch. In the coarse branch, we use the pyramidal feature hierarchy (PFH) to generate multi-scale BEV feature maps, which retain the advantages of different levels and serves as the input of the fine branch. In the fine branch, our proposed pyramid splitting and aggregation (PSA) module deeply integrates different levels of multi-scale feature maps, thereby improving the expressive ability of the final features. Extensive experiments on the challenging KITTI-3D benchmark show that our method has better performance in both 3D and BEV object detection compared with some previous state-of-the-art methods. Experimental results with average precision (AP) prove the effectiveness of our network.

Automated segmentation and evaluation of carotid plaques ultrasound images is of great significance for the diagnosis and early intervention of high-risk groups of cardiovascular and cerebrovascular diseases. However, it remains challenging to develop such solutions due to the relatively low quality of ultrasound images and heterogenous characteristics of carotid plaques. To address those problems, in this paper, we propose a novel deep convolutional neural network, FRDD-Net, with an encoder–decoder architecture to automatically segment carotid plaques. We propose the feature remapping modules (FRMs) and incorporate them into the encoding and decoding blocks to ameliorate the reliability of acquired features. We also propose a new dense decoding mechanism as part of the decoder, thus promoting the utilization efficiency of encoded features. Additionally, we construct a compound loss function to train our network to further enhance its robustness in the face of numerous cases. We train and test our network in multiple carotid plaque ultrasound datasets and our method yields the best performance compared to other state-of-the-art methods. Further ablation studies consistently show the advancement of our proposed architecture.

Automatic ship detection in optical remote sensing images is of great significance due to its broad applications in maritime security and fishery control. Most ship detection algorithms utilize a single-band image to design low-level and hand-crafted features, which are easily influenced by interference like clouds and strong waves and not robust for large-scale variation of ships. In this paper, we propose a novel coarse-to-fine ship detection method based on discrete wavelet transform (DWT) and a deep residual dense network (DRDN) to address these problems. First, multi-spectral images are adopted for sea-land segmentation, and an enhanced DWT is employed to quickly extract ship candidate regions with missing alarms as low as possible. Second, panchromatic images with clear spatial details are used for ship classification. Specifically, we propose the local residual dense block (LRDB) to fully extract semantic feature via local residual connection and densely connected convolutional layers. DRDN mainly consists of four LRDBs and is designed to further remove false alarms. Furthermore, we exploit the multiclass classification strategy, which can overcome the large intra-class difference of targets and identify ships of different sizes. Extensive experiments demonstrate that the proposed method has high robustness in complex image backgrounds and achieves higher detection accuracy than other state-of-the-art methods.

Within the development of the deep convolutional neural network, the great achievements had been made in the single-image super-resolution (SISR) task. However, the higher performance always comes with the deeper layers which also brings larger numbers of network operations and parameters that make it hard to implement in practice. In our work, a lightly super-resolution, named Mobile Share-Source Network (MSSN), is purposed to address these practical issues. In MSSN, a high-efficiency block, the mobile adaptive weighted residual unit, is designed to fulfill the need for the reduction in both parameters and the Mult-Adds while maintaining the performance with importing the deep separable convolution. Moreover, it brings into the Adaptive Weighted Share-Source Skip Connection, getting abundant information from the shallow layer which helps reconstruct better images. The experimental results show that our network has fewer numbers of parameters and operations than the state-of-the-art lightweight network while maintaining high reconstruction quality comparing with many state-of-the-art super-resolution methods in terms of both peak signal-to-noise ratio (PSNR) and structural similarity index metrics (SSIM).