No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
ResearchGate has not been able to resolve any citations for this publication.
Nowadays, autonomous driving technology has become widely prevalent. The intelligent vehicles have been equipped with various sensors (e.g. vision sensors, LiDAR, depth cameras etc.). Among them, the vision systems with tailored semantic segmentation and perception algorithms play critical roles in scene understanding. However, the traditional supervised semantic segmentation needs a large number of pixel-level manual annotations to complete model training. Although few-shot methods reduce the annotation work to some extent, they are still labor intensive. In this paper, a self-supervised few-shot semantic segmentation method based on Multi-task Learning and Dense Attention Computation (dubbed MLDAC) is proposed. The salient part of an image is split into two parts; one of them serves as the support mask for few-shot segmentation, while cross-entropy losses are calculated between the other part and the entire region with the predicted results separately as multi-task learning so as to improve the model’s generalization ability. Swin Transformer is used as our backbone to extract feature maps at different scales. These feature maps are then input to multiple levels of dense attention computation blocks to enhance pixel-level correspondence. The final prediction results are obtained through inter-scale mixing and feature skip connection. The experimental results indicate that MLDAC obtains 55.1% and 26.8% one-shot mIoU self-supervised few-shot segmentation on the PASCAL-5i and COCO-20i datasets, respectively. In addition, it achieves 78.1% on the FSS-1000 few-shot dataset, proving its efficacy.
The port-starboard ambiguity in the conventional single towed linear array sonar is one of the most deceiving obstacles which exist in the way of development of spatial spectrum estimation. A novel spatial spectrum sparse reconstruction method based on multiple beam space measurements is proposed. Firstly, the array shape of maneuvering towed array is estimated using the recursive method and the array sparse signal model is established. Then, based on target bearing variation characteristics caused by vessel's turning, the mutual incoherent property is analyzed to ensure the proposed algorithm possessing better spatial spectrum reconstruction property and the sparse reconstruction model based on continuous multiple beam snapshots is established. Simulation results demonstrate that compared with the conventional beamforming algorithm, the proposed algorithm has evident advantage in ambiguity suppression ratio and direction of arrival estimation performance. © 2017, Editorial Office of Systems Engineering and Electronics. All right reserved.
Due to extensive applications of constant false alarm rate (CFAR) algorithms to the synthetic aperture radar (SAR) image target detection, an analysis on fast CFAR detection algorithms is significant. This paper begins with a survey of fast CFAR detection algorithms. An intensive analysis on their performance and computational complexity is performed with the most two important steps, i.e., rapid pre-screening and iterative computation. The iterative computational formulae of four common CFAR detectors are then presented. A novel framework on fast CFAR detection algorithm analysis is proposed, which can be employed to analyze all of the current fast algorithms. Finally, based on the classic two-parameter CFAR algorithm, the framework is simulated and analyzed, and its feasibility and performance are validated. The results show that the new framework sufficiently integrates the advantages of rapid pre-screening and iterative computation, and effectively improves execution efficiency of the CFAR algorithm in SAR images detection.
A rolling bearing fault diagnosis method based on MTF-MSMCNN with small sample[J]
- DUAN
Application of interpolation method based on box plots in data processing of aneroid barometer[J]
- WU
A ship detection algorithm for SAR image based on box-plot[J]
- Zhao
Research Review for Estimating Technique of Towed Line Array Shape[J]
- ZHANG
Review on the development of foreign anti-submarine detection equipment and technology
- YANG
An Optimization Method of Underwater Acoustic Target Classification Model Based on Sample Expansion Network[J]
- ZHANG
Piston pump fault diagnosis based on Siamese neural network with small samples[J]
- GAO
Research on Underwater Acoustic Array Signal Processing Technology Based on Compressive Sensing
- Qidong