融合深度网络的改进快速生成超像素算法

doi:10.3778/j.issn.1673-9418.2001021

摘要/Abstract

摘要：

超像素是图像过度分割的结果，提供了图像数据的中间级表示，对计算机视觉等领域的研究具有重要意义。现有的超像素算法是不可微的，且深度网络通常在规则的网格上进行定义，导致目前生成超像素的算法大多基于手工提取的像素特征进行。提出融合深度网络的改进快速生成超像素算法，将深度学习网络嵌入到超像素的生成过程中，首先利用含多隐含层的深度网络进行图像像素特征的提取，然后通过K-means聚类方法计算初始种子点位置以改善分割结果，在此基础上通过主动搜索方法确保像素标签的正确性，最后得到超像素分割结果。在Berkeley数据集BSDS500上，使用BSDS的基准测评与其他文献的对比实验表明，所提出的算法在分割结果的紧凑性、规则性等性能方面相对较好。

关键词: 超像素分割, 深度网络, 聚类, 特征提取

Abstract:

Superpixels are the result of over-segmentation of the image and provide an intermediate representation of the image data. It plays an important role in the research of computer vision and other fields. However, the existing superpixel algorithms are non-differentiable and the deep networks are usually defined over regular grid. Thus, most of the current superpixel generation algorithms are based on hand-crafted pixel features. To this end, this paper proposes an improved fast generation of superpixel algorithms with deep network. Firstly, the deep network with multiple hidden layers is directly embedded in the process of superpixel generation to extract the pixel features of the image. Secondly, this paper calculates the initial seed point position by [K-means] clustering method to signi-ficantly improve the segmentation result, and an active search method is used to ensure the correctness of the pixel label. Finally, superpixel segmentation results are obtained. On the Berkeley dataset BSDS500, the benchmark eva-luation using BSDS is compared with other literatures, and the experimental results show that the proposed algo-rithm performs relatively well in terms of the compactness and regularity of segmentation results.

Key words: superpixel segmentation, deep network, clustering, feature extraction

盛家川，王佳媛，李玉芝，王君. 融合深度网络的改进快速生成超像素算法[J]. 计算机科学与探索, 2020, 14(12): 2132-2139.

SHENG Jiachuan, WANG Jiayuan, LI Yuzhi, WANG Jun. Improved Fast Generation of Superpixel Algorithms with Deep Network[J]. Journal of Frontiers of Computer Science and Technology, 2020, 14(12): 2132-2139.

参考文献

[1] Sheng J C, Li Y Z. Learning artistic objects for improved classi-fication of Chinese paintings[J]. Journal of Image and Graphics, 2018, 23(8): 1193-1206.盛家川, 李玉芝. 国画的艺术目标分割及深度学习与分类[J]. 中国图象图形学报, 2018, 23(8): 1193-1206.
[2] Zhang Y Y, Ge H W. Image saliency detection combining sparse reconstruction and compactness[J]. Journal of Frontiers of Computer Science and Technology, 2020, 14(12): 2122-2131.张莹莹, 葛洪伟. 稀疏重构和紧凑性结合的图像显著性检测[J]. 计算机科学与探索, 2020, 14(12): 2122-2131.
[3] Gui Y, Tang W, Zeng G. Gradient-constrained SLIC based fast video object segmentation[J]. Journal of Frontiers of Computer Science and Technology, 2019, 13(2): 285-299. 桂彦, 汤问, 曾光. 梯度约束SLIC的快速视频目标提取方法[J]. 计算机科学与探索, 2019, 13(2): 285-299.
[4] Xue S, Wang W J. Depth estimation from single defocused image based on superpixel segmentation[J]. Journal of Frontiers of Computer Science and Technology, 2018, 12(7): 1162-1168.薛松, 王文剑. 基于超像素分割的单幅散焦图像深度恢复方法[J]. 计算机科学与探索, 2018, 12(7): 1162-1168.
[5] Achanta R, Shaji A, Smith K, et al. SLIC superpixels compared to state-of-the-art superpixel methods[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2012, 34(11): 2274-2282.
[6] Lloyd S P. Least squares quantization in PCM[J]. IEEE Trans-actions on Information Theory, 1982, 28(2): 129-137.
[7] Li Z Q, Chen J S. Superpixel segmentation using linear spectral clustering[C]//Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition, Boston, Jun 7-12, 2015. Washington: IEEE Computer Society, 2015: 1356-1363.
[8] Yang Y, Xu D Y. SLIC algorithm based on optimizing initial center point of weighted kernel K-means clustering[J]. Journal of Frontiers of Computer Science and Technology, 2018, 12(3): 494-501.杨艳, 许道云. 优化加权核K-means聚类初始中心点的SLIC算法[J]. 计算机科学与探索, 2018, 12(3): 494-501.
[9] Wang P, Zeng G, Gan R, et al. Structure-sensitive super-pixels via geodesic distance[J]. International Journal of Com-puter Vision, 2013, 103(1): 1-21.
[10] Liu Y J, Yu C C, Yu M J, et al. Manifold SLIC: a fast method to compute content-sensitive superpixels[C]//Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, Jun 27-30, 2016. Washington: IEEE Computer Society, 2016: 651-659.
[11] Zhao J X, Ren B, Hou Q B, et al. FLIC: fast linear iterative clustering with active search[C]//Proceedings of the 32nd AAAI Conference on Artificial Intelligence, the 30th Innov-ative Applications of Artificial Intelligence, and the 8th AAAI Symposium on Educational Advances in Artificial Intelligence, New Orleans, Feb 2-7, 2018. Menlo Park: AAAI, 2018: 7574-7581.
[12] Sheng J C, Chen Y Q, Wang J, et al. Chinese paintings senti-ment recognition via CNN optimization with human cognition[J]. Pattern Recognition and Artificial Intelligence, 2020, 33(2): 141-149.盛家川, 陈雅琦, 王君, 等. 融合人类认知网络优化的中国画情感识别[J]. 模式识别与人工智能, 2020, 33(2): 141-149.
[13] Weimer D, Scholz-Reiter B, Shpitalni M. Design of deep convolutional neural network architectures for automated feature extraction in industrial inspection[J]. CIRP Annals, 2016, 65(1): 417-420.
[14] Zhang W, Zheng K, Tang P, et al. Land cover classification with features extracted by deep convolutional neural net-work[J]. Journal of Image and Graphics, 2017, 22(8): 1144-1153.张伟, 郑柯, 唐娉, 等. 深度卷积神经网络特征提取用于地表覆盖分类初探[J]. 中国图象图形学报, 2017, 22(8): 1144-1153.
[15] Jampani V, Sun D Q, Liu M Y, et al. Superpixel sampling networks[C]//LNCS 11211: Proceedings of the 15th European Conference on Computer Vision, Munich, Sep 8-14, 2018. Berlin, Heidelberg: Springer, 2018: 363-380.
[16] Gadde R, Jampani V, Kiefel M, et al. Superpixel convo-lutional networks using bilateral inceptions[C]//LNCS 9905: Proceedings of the 14th European Conference on Computer Vision, Amsterdam, Oct 11-14, 2016. Berlin, Heidelberg:Springer, 2016: 597-613.
[17] He S, Lau R W H, Liu W, et al. SuperCNN: a superpixel-wise convolutional neural network for salient object detec-tion[J]. International Journal of Computer Vision, 2015, 115(3): 330-344.
[18] Sharma A, Tuzel O, Liu M Y. Recursive context propaga-tion network for semantic scene labeling[J]. Advances in Neural Information Processing Systems, 2014, 3: 2447-2455.
[19] Arbelaez P A, Maire M, Fowlkes C C, et al. Contour detec-tion and hierarchical image segmentation[J]. IEEE Transac-tions on Pattern Analysis and Machine Intelligence, 2011, 33(5): 898-916.
[20] Kingma D P, Ba J. Adam: a method for stochastic optimi-zation[C]//Proceedings of the 3rd International Conference on Learning Representations, San Diego, May 7-9, 2015: 1-15.
[21] Barnes C, Shechtman E, Finkelstein A, et al. PatchMatch: a randomized correspondence algorithm for structural image editing[J]. ACM Transactions on Graphics, 2009, 28(3): 24.