As an important task in scene understanding, semantic segmentation requires a large amount of computation to achieve high performance. In recent years, with the rise of autonomous systems, it is crucial to make a trad...
详细信息
As an important task in scene understanding, semantic segmentation requires a large amount of computation to achieve high performance. In recent years, with the rise of autonomous systems, it is crucial to make a trade-off in terms of accuracy and speed. In this paper, we propose a novel asymmetric encoder-decoder network structure to address this problem. In the encoder, we design a Separable Asymmetric Module, which combines depth-wise separable asymmetric convolution with dilated convolution to greatly reduce computation cost while maintaining accuracy. On the other hand, an attention mechanism is also used in the decoder to further improve segmentation performance. Experimental results on CityScapes and CamVid datasets show that the proposed method can achieve a better balance between segmentation precision and speed compared with state-of-the-art semantic segmentation methods. Specifically, our model obtains mean IoU of 72.5% and 66.3% on CityScapes and CamVid test dataset, respectively, with less than 1M parameters.
A switchable-rate quasi-cyclic low-density parity-check (QC-LDPC) coding scheme has been proposed for integration within the legacy and next-generation high-frequency internet protocol (HF-IP) systems. The novelty in ...
详细信息
A switchable-rate quasi-cyclic low-density parity-check (QC-LDPC) coding scheme has been proposed for integration within the legacy and next-generation high-frequency internet protocol (HF-IP) systems. The novelty in this work is based upon using a class of switchable-rate short-block-length (<1500 bits) QC-LDPC codes for the HF fading channel modelled by the ITU-R F. 1487 for all latitudes and conditions. The QC-LDPC codes are constructed using a switchable-rate approach based on finite fields which provides the ability to switch among three rates to combat varying channel conditions using a single encoder/decoder structure. The proposed structure enables low-complexity implementation of the low-density parity-check encoder/decoder for use within the existing data link (DL) layer of the standardisation agreement (STANAG) 5066 profile. The performance of the proposed scheme has been evaluated comprehensively for all the HF channel conditions and latitudes. A comparison between the proposed and the current coding scheme in HF-IP systems (based on convolutional coding) shows an improvement in error-rate performance.
Quasi-cyclic (QC) low-density parity-check (LDPC) codes have the parity-check matrices consisting of circulant matrices. Since QC LDPC codes whose parity-check matrices consist of only circulant permutation matrices a...
详细信息
Quasi-cyclic (QC) low-density parity-check (LDPC) codes have the parity-check matrices consisting of circulant matrices. Since QC LDPC codes whose parity-check matrices consist of only circulant permutation matrices are difficult to support layered decoding and, at the same time, have a good degree distribution with respect to error correcting performance, adopting multi-weight circulant matrices to parity-check matrices is useful but it has not been much researched. In this paper, we propose a new code structure for QC LDPC codes with multi-weight circulant matrices by introducing overlapping matrices. This structure enables a system to operate on dual mode in an efficient manner, that is, a standard QC LDPC code is used when the channel is relatively good and an enhanced QC LDPC code adopting an overlapping matrix is used otherwise. We also propose a new dual mode parallel decoder which supports the layered decoding both for the standard QC LDPC codes and the enhanced QC LDPC codes. Simulation results show that QC LDPC codes with the proposed structure have considerably improved error correcting performance and decoding throughput. (c) 2013 Elsevier GmbH. All rights reserved.
Real-time monitoring and surveillance play an important role in the field of remote sensing, where multi-spectral (MS) images with high spatial resolution are widely desired for better analysis. However, high-resoluti...
详细信息
Real-time monitoring and surveillance play an important role in the field of remote sensing, where multi-spectral (MS) images with high spatial resolution are widely desired for better analysis. However, high-resolution MS images cannot be directly obtained due to the limitations of sensors and bandwidth. As an essential way to alleviate this problem, pan-sharpening aims at fusing the complementary information of a low-resolution MS image and a high-resolution panchromatic (PAN) image to reconstruct a high-resolution MS image. Most previous deep-learning based methods can meet the real-time requirements with the help of graphics processing unit (GPU). However, they don't fully exploit the favorable hierarchical information, sparing huge room for performance improvement. In this paper, to meet the requirement of real-time implementation and achieve more effective performance simultaneously, we propose a multi-scale fusion network (MSFN) to make full use of hierarchical complementary features of PAN and MS images. Specifically, we introduce an encoder-decoder structure and coarse-to-fine strategy to effectively extract multi-scale features of PAN and MS images, separately. Meanwhile, an information pool is adopted to preserve primitive information. Then a multi-scale feature fusion module is applied to fuse multi-scale features from the decoder and information pool. Finally, the fused features are utilized to reconstruct the high-resolution MS image. Extensive experiments demonstrate that our proposed method achieves favorable performance against other methods in terms of quantitative metrics and visual quality. Besides, the results on running time indicate that our method can achieve real-time performance.
Image semantic segmentation has been applied more and more widely in the fields of satellite remote sensing, medical treatment, intelligent transportation, and virtual reality. However, in the medical field, the study...
详细信息
Image semantic segmentation has been applied more and more widely in the fields of satellite remote sensing, medical treatment, intelligent transportation, and virtual reality. However, in the medical field, the study of cerebral vessel and cranial nerve segmentation based on true-color medical images is in urgent need and has good research and development prospects. We have extended the current state-of-the-art semantic-segmentation network DeepLabv3+ and used it as the basic framework. First, the feature distillation block (FDB) was introduced into the encoder structure to refine the extracted features. In addition, the atrous spatial pyramid pooling (ASPP) module was added to the decoder structure to enhance the retention of feature and boundary information. The proposed model was trained by fine tuning and optimizing the relevant parameters. Experimental results show that the encoder structure has better performance in feature refinement processing, improving target boundary segmentation precision, and retaining more feature information. Our method has a segmentation accuracy of 75.73%, which is 3% better than DeepLabv3+.
暂无评论