Two-level attention and score consistency network for plant segmentation

Two-level attention and score consistency network for plant segmentation
复制标题

用于植物分割的两级注意力和评分一致性网络

DOI:
10.1016/j.compag.2020.105281
复制
发表时间:
2020-03
影响因子:
8.3
通讯作者:
Lili Guo
Lili Guo
中科院分区:
农林科学1区
文献类型:
--
作者:
Lele Xu;Ye Li;Jinzhong Xu;Lili Guo

文献摘要

参考文献

相似文献

植物图像的语义分割可以为植物表型研究提供有价值的信息。然而,当面对复杂的背景干扰和不均匀的照明时,存在挑战。在这项研究中,我们提出了一个两级的注意力和分数一致性为基础的网络(TASCN)的植物语义分割。我们的TASCN包括一个两级注意力子模块(TAM)和一个多尺度特征融合子模块(FFM)。TAM结合了自上而下的语义注意门(层间注意,级别1)和自注意机制(层内注意,级别2)来突出显着特征并抑制不相关或干扰信息。FFM融合TAM的多尺度特征,得到最终的分割分数图。通过分数一致性损失对分数图进行进一步约束,提高了语义一致性,降低了光照不均匀的影响。在中国空间实验室TG-2的水稻数据集上的实验表明,我们的TASCN分割性能令人印象深刻,特别是对复杂的背景干扰和不均匀的照明。此外,定量评估(平均IoU)和定性可视化结果都表明,我们的TASCN优于最先进的方法,包括FCN,U-Net,SegNet,PSPNet,DeepLabv 3+和DANet。总之,提出了一种新型的TASCN,以满足植物研究中日益增长的精细植物解析需求。
Semantic segmentation of plant images can provide valuable information for plant phenotypic studies. However, challenges exist when facing complex background interference and uneven illumination. In this study, we propose a Two-level Attention and Score Consistency based Network (TASCN) for semantic segmentation of plant. Our TASCN includes a two-level attention sub-module (TAM) and a multi-scale feature fusion sub-module (FFM). The TAM combines a top-down semantic attention gate (Inter-layer attention, level 1) and a self-attention mechanism (Intra-layer attention, level 2) to highlight the salient features and suppress the irrelevant or interference information. The FFM fuses the multi-scale features from the TAM to obtain the final segmentation score-map. The score-map is further restricted with the score consistency loss to promote the semantic consistency and reduces the impact of uneven illumination. Experiments on rice dataset from the Chinese spacelab TG-2 demonstrate the impressing segmentation performance of our TASCN, especially for complex background interference and uneven illumination. In addition, both the quantitative evaluation (Mean IoU) and qualitative visual results show that our TASCN outperforms the state-of-the-art methods, including FCN, U-Net, SegNet, PSPNet, DeepLabv3+ and DANet. In conclusion, a novel TASCN is proposed to meet the increasing need for fine plant parsing in plant research.
TG-2空间实验室微重力条件下水稻幼苗光周期控制的发育与生长
DOI: 10.1007/s12217-018-9644-3
发表时间: 2018
影响因子: 1.8
作者:
Lihua Wang;Fei Han;Huiqiong Zheng
通讯作者: Huiqiong Zheng
DOI: 10.1109/icra.2018.8460962
发表时间: 2017-09
期刊: 2018 IEEE International Conference on Robotics and Automation (ICRA)
影响因子: --
作者:
Andres Milioto;Philipp Lottes;C. Stachniss
通讯作者: Andres Milioto;Philipp Lottes;C. Stachniss
DOI: 10.1007/978-3-642-38797-5_7
发表时间: 2014
期刊: --
影响因子: --
作者:
S. Singh
通讯作者: S. Singh
DOI: 10.1016/j.patcog.2007.07.007
发表时间: 2008-02
期刊: Pattern Recognit.
影响因子: --
作者:
A. Tellaeche;X. Burgos-Artizzu;G. Pajares;A. Ribeiro
通讯作者: A. Tellaeche;X. Burgos-Artizzu;G. Pajares;A. Ribeiro
DOI: 10.1109/tgrs.2018.2866056
发表时间: 2019-03-01
影响因子: 8.2
作者:
Jin, Shichao;Su, Yanjun;Guo, Qinghua
通讯作者: Guo, Qinghua