[author_cn_name].[cn_title][J].空军工程大学学报:自然科学版,[year_id],[volume]([issue]):[start_page]-[end_page] 一种不同色域空间下的无监督图像分割技术- A Kind of Unsupervised Segmentation Technique in Different Color Space
文章摘要
吴涛, 王伦武, 王伦文, 朱敬成.一种不同色域空间下的无监督图像分割技术[J].空军工程大学学报:自然科学版,2022,23(1):104-111
一种不同色域空间下的无监督图像分割技术
A Kind of Unsupervised Segmentation Technique in Different Color Space
  
DOI:
中文关键词: 色域空间  超像素  图像分割  无监督学习
英文关键词: color space  super pixel  image segmentation  unsupervised learning
基金项目:
作者单位
吴涛, 王伦武, 王伦文, 朱敬成 国防科技大学电子对抗学院合肥230037 
摘要点击次数: 87
全文下载次数: 104
中文摘要:
      针对图像分割有监督学习模型训练时间长、需要大量训练样本才能确保模型精度要求且样本标记费时费力的问题,提出了在不同色域空间下基于神经网络的无监督图像分割方法。首先将图像进行不同颜色空间模型转化,得到不同色域空间下图像的颜色表示;其次利用felz和quickshift方法,对转换后的图像进行粗粒度聚类,形成超像素结果,并对每个像素打上相应的标签;最后利用神经网络细粒度的图像特征分辨能力进行微调,得到最终的图像分割结果。该方法在公开的COD10K等数据集上选取了数据集进行验证,实验表明,文中方法能够对图像进行合理分割,且与有监督长时间训练过程相比,无监督的推理耗时大大缩短,速度显著提高。
英文摘要:
      Aimed at the problems that the supervised learning model for image segmentation is long in training time, a large number of training samples is needed to ensure model accuracy requirement, and the labeling takes a lot of work, time and energy, an unsupervised image segmentation method is proposed based on neural network in different color space. Firstly, the images are transformed into different color space models to obtain the color representation of images in different color gamut spaces, and then by using felz and quickshift methods, a coarse grained cluster is made after the images being transformed to form super pixel results and label each pixel accordingly. Finally, the fine grained image feature discrimination ability of neural networks is utilized for making a fine turning, obtaining the final image segmentation results. The method is validated on publicly available datasets such as COD10K selected, and the experiments prove that the proposed method can segment images reasonably with less inference time consumption and faster speed in comparison with the supervised training for a long time.
查看全文   查看/发表评论  下载PDF阅读器
关闭