SKETCH-BASED IMAGE RETRIEVAL VIA CAT LOSS WITH ELASTIC NET REGULARIZATION

SKETCH-BASED IMAGE RETRIEVAL VIA CAT LOSS WITH ELASTIC NET REGULARIZATION
复制标题

通过具有弹性网络正则化的猫丢失进行基于草图的图像检索

DOI:
10.3934/mfc.2020013
复制
发表时间:
2020
影响因子:
1.4
通讯作者:
Hu Zhensheng
Hu Zhensheng
中科院分区:
--
文献类型:
--
作者:
Cai Jia;Xu Guanglong;Hu Zhensheng

文献摘要

参考文献

相似文献

基于细粒度草图的图像检索(FG-SBIR)是一个重要的问题,它使用手绘草图作为查询来执行照片的实例级检索。人类草图通常是高度抽象和标志性的,这使得FG-SBIR是一项具有挑战性的任务。现有的FG-SBIR方法采用l(2)正则化三重态损失或高阶能量函数进行检索,忽略了不同域(草图、照片)之间的特征间隙,需要选择权层矩阵。这产生了很高的计算复杂度。本文定义了一种基于注意力模型的弹性网正则化CAT损失函数。它可以缩小不同子网之间的特征差距,体现草图的稀疏性。实验表明,所提出的方法与最先进的方法相比具有竞争力。
Fine-grained sketch-based image retrieval (FG-SBIR) is an important problem that uses free-hand human sketch as queries to perform instance-level retrieval of photos. Human sketches are generally highly abstract and iconic, which makes FG-SBIR a challenging task. Existing FG-SBIR approaches using triplet loss with l(2) regularization or higher-order energy function to conduct retrieval performance, which neglect the feature gap between different domains (sketches, photos) and need to select the weight layer matrix. This yields high computational complexity. In this paper, we define a new CAT loss function with elastic net regularization based on attention model. It can close the feature gap between different subnetworks and embody the sparsity of the sketches. Experiments demonstrate that the proposed approach is competitive with state-of-the-art methods.
DOI: 10.1109/cvpr.2017.247
发表时间: 2017-03
期刊: 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
影响因子: --
作者:
Li Liu;Fumin Shen;Yuming Shen;Xianglong Liu;Ling Shao
通讯作者: Li Liu;Fumin Shen;Yuming Shen;Xianglong Liu;Ling Shao
DOI: 10.1109/tpami.2010.180
发表时间: 2011-03-01
影响因子: 23.6
作者:
Klare, Brendan F.;Li, Zhifeng;Jain, Anil K.
通讯作者: Jain, Anil K.
DOI: 10.1145/2661229.2661231
发表时间: 2014-11-01
影响因子: 6.2
作者:
Schneider, Rosalia G.;Tuytelaars, Tinne
通讯作者: Tuytelaars, Tinne
DOI: 10.1007/978-3-319-59876-5_33
发表时间: 2017-07
期刊: --
影响因子: --
作者:
Omar Seddati;S. Dupont;S. Mahmoudi
通讯作者: Omar Seddati;S. Dupont;S. Mahmoudi
DOI: 10.1109/vcip.2017.8305153
发表时间: 2017-12
期刊: 2017 IEEE Visual Communications and Image Processing (VCIP)
影响因子: --
作者:
H. Zhang;Chuang Zhang;Ming Wu
通讯作者: H. Zhang;Chuang Zhang;Ming Wu