| 注意力残差多尺度特征增强的显著性实例分割 |
| |
| 引用本文： | 史彩娟,陈厚儒,葛录录,王子雯. 注意力残差多尺度特征增强的显著性实例分割[J]. 图学学报, 2021, 42(6): 883-890. DOI: 10.11996/JG.j.2095-302X.2021060883 |
| |
| 作者姓名： | 史彩娟 陈厚儒 葛录录 王子雯 |
| |
| 作者单位： | 华北理工大学人工智能学院，河北 唐山 063210 |
| |
| 基金项目： | 国家自然科学基金项目(61502143)；河北省研究生示范课项目(KCJSX2019097)；华北理工大学杰出青年基金项目(JQ201715)；唐山市人才资助项目(A202110011) |
| |
| 摘 要： | 显著性实例分割是指分割出图像中最引人注目的实例对象。现有的显著性实例分割方法中存在较小显著性实例不易检测分割，以及较大显著性实例分割精度不足等问题。针对这 2 个问题，提出了一种新的显著性实例分割模型，即注意力残差多尺度特征增强网络(ARMFE)。模型 ARMFE 主要包括 2 个模块：注意力残差网络模块和多尺度特征增强模块，注意力残差网络模块是在残差网络基础上引入注意力机制，分别从通道和空间对特征进行选择增强；多尺度特征增强模块则是在特征金字塔基础上进一步增强尺度跨度较大的特征信息融合。因此，ARMFE 模型通过注意力残差多尺度特征增强，充分利用多个尺度特征的互补信息，同时提升较大显著性实例对象和较小显著性实例对象的分割效果。ARMFE 模型在显著性实例分割数据集 Salient InstanceSaliency-1K (SIS-1K)上进行了实验，分割精度和速度都得到了提升，优于现有的显著性实例分割算法 MSRNet和 S4Net。
|
| 关 键 词： | 显著性实例分割 注意力机制 残差网络 多尺度 特征增强 |
Salient instance segmentation via attention residual multi-scale feature enhancement |
| |
| SHI Cai-juan,CHEN Hou-ru,GE Lu-lu,WANG Zi-wen. Salient instance segmentation via attention residual multi-scale feature enhancement[J]. Journal of Graphics, 2021, 42(6): 883-890. DOI: 10.11996/JG.j.2095-302X.2021060883 |
| |
| Authors: | SHI Cai-juan CHEN Hou-ru GE Lu-lu WANG Zi-wen |
| |
| Affiliation: | College of Artificial Intelligence, North China University of Science and Technology, Tangshan Hebei 063210, China |
| |
| Abstract: | Salient instance segmentation is to segment the most noticeable instance object in the image. However, thereremain some problems in the existing methods of salient instance segmentation. For example, the small salient instancesare difficult to be detected and segmented, and the segmentation accuracy is insufficient for large salient instances.Therefore, to solve these two problems, a new salient instance segmentation model, namely the attention residualmulti-scale feature enhancement network (ARMFE), has been proposed. ARMFE includes two modules, i.e. the attentionresidual network module and the multi-scale feature enhancement module. The attention residual network modulecombines the residual network with the spatial attention sub-module and the channel attention sub-module to enhance thefeatures. The multi-scale feature enhancement module can further enhance the information fusion for features with largescale span based on the feature pyramid. Therefore, the proposed ARMFE model can make full use of thecomplementary information of multi-scales features by attention residual multi-scale feature enhancement, and then simultaneously improve the accuracy of detecting and segmenting large instance objects and small instance objects. Theproposed ARMFE model has been tested on the salient instance segmentation dataset Salient Instance Saliency-1K(SIS-1K), and the segmentation accuracy and speed have been improved. This indicates that our proposed modeloutperforms other existing salient instance segmentation algorithms, such as MSRNet and S4Net. |
| |
| Keywords: | salient instance segmentation attention mechanism residual network multi-scale feature enhancement |
| 本文献已被 万方数据 等数据库收录！ |
| 点击此处可从《图学学报》浏览原始摘要信息 |
|
点击此处可从《图学学报》下载免费的PDF全文 |
