ZHAO Yi,HE Jia.Garment Image Segmentationn Using Dual Attention Mechanism Deeplabv3+ Algorithm[J].Journal of Chengdu University of Information Technology,2022,37(01):67-71.[doi:10.16836/j.cnki.jcuit.2022.01.012]
采用双注意力机制Deeplabv3+算法的服装图像分割
- Title:
- Garment Image Segmentationn Using Dual Attention Mechanism Deeplabv3+ Algorithm
- 文章编号:
- 2096-1618(2022)01-0067-05
- 关键词:
- 服装图像分割; DeepFashion2; Deeplabv3+; 语义分割
- 分类号:
- TP391.41
- 文献标志码:
- A
- 摘要:
- 近年来服装时尚行业经济发展迅速，为了让用户选择服装和服装的设计更方便快捷，提高服装图像的分割效率尤为重要。目前的方法大多属于传统的分割方法，或者基于深度卷积神经网络(DCNN)。针对服装图像分割时易受背景、颜色、纹理等的影响，且服装的边缘分割不准确，基于Deeplabv3+算法提出了双注意力机制的方法识别分割服装图像，使用通道注意力机制和位置注意力机制构成名为CPAM的模块对Deeplabv3+网络进行改进。特征图经过多次下采样后再经过通道和位置注意力模块(CPAM)与ASPP模块并行，最后通过上采样得到预测图像。实验证明对不同场景的服装图像分割，加入CPAM模块的模型能更准确地将服装分割出来。
- Abstract:
- In recent years, the garment fashion industry economy has developed rapidly. In order to make the user’s choice of clothing and clothing design more convenient and fast, it is particularly important to improve the efficiency of clothing segmentation. Most of the present methods are traditional segmentation methods or based on deep convolutional neural network(DCNN). For the clothing image segmentation task is easily affected by background, color, texture, etc., and the clothing edge segmentation is not accurate, this paper proposes a method of dual-attention mechanism based Deeplabv3+ algorithm to identify and segment clothing images. Channel attention mechanism and location attention mechanism are used to form a module named CPAM to improve Deeplabv3+ network. After downsampling for several times, the feature image is parallel to the channel and position attention module(CPAM)and The ASPP module, and then the prediction image is obtained by upsampling. Finally, the experiment proves that the model with CPAM module can segment the clothing image more accurately in different scenes.
参考文献/References:
[1] Jouanneau W,Bugeau A,Palyart M,et al.Where Are My Clothes? A Multi-Level Approach for Evaluating Deep Instance Segmentation Architectures on Fashion Images[C].Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition,2021:3951-3955.
[2] Castro H,Ramirez M.Segmentation task for fashion and apparel[EB/OL].arXiv preprint arXiv:2006:11375.
[3] 张艳红,杨思,徐增波.图像分割技术在服装领域的应用[J].软件导刊,2020,19(4):238-241.
[4] Inacio A D S,Lopes H S.EPYNET: Efficient Pyramidal Network for Clothing Segmentation[J].IEEE Access,2020,8:187882-187892.
[5] Silvestre L.Regularity of the obstacle problem for a fractional power of the Laplace operator[J].Communications on Pure and Applied Mathematics: A Journal Issued by the Courant Institute of Mathematical Sciences,2007,60(1):67-112.
[6] 黄冬艳,刘骊,付晓东,等.基于HOG和E-SVM的服装图像联合分割算法[J].计算机工程与应用,2017,53(18):199-203.
[7] 李冬艳,陈文雄.上半空间高次分数阶Laplace方程解的不存在性[J].纺织高校基础科学学报,2017,30(1):18-22.
[8] 白美丽,万韬阮,汤汶,等.一种改进的用于服装解析的自监督网络学习方法[J].纺织高校基础科学学报,2019,32(4):385-392.
[9] Krizhevsky A,Sutskever I,Hinton G E. Imagenet classification with deep convolutional neural networks[J].Advances in neural information processing systems,2012,25:1097-1105.
[10] Sun Y,Wang X,Tang X.Deep convolutional network cascade for facial point detection[C].Proceedings of the IEEE conference on computer vision and pattern recognition,2013:3476-3483.
[11] Long J,Shelhamer E,Darrell T.Fully convolutional networks for semantic segmentation[C].Proceedings of the IEEE conference on computer vision and pattern recognition,2015:3431-3440.
[12] Chen L C,Zhu Y,Papandreou G,et al.Encoder-decoder with atrous separable convolution for semantic image segmentation[C].Proceedings of the European conference on computer vision(ECCV),2018:801-818.
[13] 王中宇,倪显扬,尚振东.利用卷积神经网络的自动驾驶场景语义分割[J].光学精密工程,2019,27(11):2429-2438.
[14] Hu J,Shen L,Sun G.Squeeze-and-excitation networks[C].Proceedings of the IEEE conference on computer vision and pattern recognition,2018:7132-7141.
[15] He K,Zhang X,Ren S,et al.Spatial pyramid pooling in deep convolutional networks for visual recognition[J].IEEE transactions on pattern analysis and machine intelligence,2015,37(9):1904-1916.
[16] Ge Y,Zhang R,Wang X,et al.Deepfashion2:A versatile benchmark for detection, pose estimation,segmentation and re-identification of clothing images[C].Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition,2019:5337-5345.
[17] Yamaguchi K,Hadi Kiapour M,Berg T L.Paper doll parsing: Retrieving similar styles to parse clothing items[C].Proceedings of the IEEE international conference on computer vision,2013:3519-3526.
备注/Memo
收稿日期:2021-10-14