JOURNAL ARTICLE
基于改进Deeplab V3+网络的语义分割
Abstract
深度学习的语义分割在计算机视觉领域中有非常广阔的发展前景, 但许多分割效果较好网络模型占用内存大和处理单张图片耗时长. 针对这个问题, 把Deeplab V3+模型的骨干网(ResNet101)的瓶颈单元设计为1D非瓶颈单元, 且对空洞空间金字塔池化模块(Atrous Spatial Pyramid Pooling, ASPP)的卷积层进行分解. 该算法能大幅度降低Deeplab V3+网络的参数量, 提高网络推理速度. 基于PASCAL VOC 2012数据集进行对比实验, 实验结果显示改进网络模型拥有更快的处理速度和更优的分割效果, 且消耗更少的内存.
Keywords:
Mathematics
Metrics
0
Cited By
0.00
FWCI (Field Weighted Citation Impact)
0
Refs
0.62
Citation Normalized Percentile
Is in top 1%
Is in top 10%
Topics
Image Processing Techniques and Applications
Physical Sciences → Engineering → Media Technology
Advanced Image and Video Retrieval Techniques
Physical Sciences → Computer Science → Computer Vision and Pattern Recognition
Brain Tumor Detection and Classification
Life Sciences → Neuroscience → Neurology
Related Documents
JOURNAL ARTICLE
基于改进DeepLabv3+网络的图像语义分割方法
Journal: Laser & Optoelectronics Progress Year: 2021 Vol: 58 (16)Pages: 1610008-1610008
JOURNAL ARTICLE
基于改进DeepLabv3+网络的轻量级语义分割算法
姚燕 Yao Yan胡立坤 Hu Likun郭军 Guo Jun
Journal: Laser & Optoelectronics Progress Year: 2022 Vol: 59 (4)Pages: 0410015-0410015
JOURNAL ARTICLE
基于改进DeepLab V3+的语义地图构建
李琳 Li Lin吴怀宇 Wu Huaiyu张天宇 Zhang Tianyu
Journal: Laser & Optoelectronics Progress Year: 2022 Vol: 59 (10)Pages: 1015002-1015002
JOURNAL ARTICLE
基于Deeplab v3+的图像语义分割算法优化研究
孟俊熙 Meng Junxi张莉 Zhang Li曹洋 Cao Yang张乐天 Zhang Letian宋倩 Song Qian
Journal: Laser & Optoelectronics Progress Year: 2022 Vol: 59 (16)Pages: 1610009-1610009
JOURNAL ARTICLE
DECANet：基于改进DeepLabv3+的图像语义分割方法
唐璐 Tang LuLiang Wan王婷婷 Wang Tingting李树胜 Li Shusheng
Journal: Laser & Optoelectronics Progress Year: 2023 Vol: 60 (4)Pages: 0410002-0410002