基于编解码和局部增强的光电图像分割算法

针对光电图像语义分割问题, 提出了一种基于编解码(Encoder-Decoder)结构和图像局部增强的分割算法。首先, 采用基于互质因子的空洞空间金字塔池化(Atrous Spatial Pyramid Pooling, ASPP)模块减小多尺度空洞卷积(Atrous Convolution)引入的网格效应, 提升卷积核的像素近邻信息表征能力; 其次, 对分割难度较大的图像局部区域, 采用融合平均交并比(Mean Intersection Over Union, MIOU)和交叉信息熵的损失函数, 结合权值衰减策略, 提高这些局部区域的像素权重。实验结果表明, 提出的改进算法能有效提升图像语义分割精度。

Abstract

For semantic segmentation of optoelectric images, it proposes a segmentation algorithm based on encoder-decoder structure and local image enhancement. Firstly, the algorithm adopts an atrous spatial pyramid pooling (ASPP) module based on coprime factors to reduce the grid effect caused by multiscale convolution and improve the capability of representing pixel adjacent information of convolution kernels. Secondly, in order to improve pixel weight of areas that are difficult to segment, the algorithm combines loss function consisting of mean intersection and cross entropy with weight decay strategy. Experimental results show that the proposed algorithm can effectively improve the accuracy of image semantic segmentation.

参考文献

[1] 魏云超, 赵耀. 基于DCNN的图像语义分割综述［J］. 北京交通大学学报, 2016, 40(4): 82-91.

Wei Yunchao, Zhao Yao. A review on image semantic segmentation based on DCNN［J］. J. of Beijing Jiaotong University, 2016, 40(4): 82-91.

[2] 欧阳鑫玉, 赵楠楠, 宋蕾, 等. 图像分割技术的发展［J］. 辽宁科技大学学报, 2002, 25(5): 363-368.

Ouyang Xinyu, Zhao Nannan, Song Lei, et al. Survey on image segmentation［J］. J. of University of Science and Technol. Liaoning, 2002, 25(5): 363-368.

[3] 代少升, 申娇, 陈桂芳, 等. 基于多阈值分割的红外图像伪彩增强算法［J］. 半导体光电, 2015, 36(5): 800-803.

Dai Shaosheng, Shen Jiao, Chen Guifang, et al. Pseudo-color enhancement algorithm of infrared images based on multi-threshold segmentation［J］. Semiconductor Optoelectronics, 2015, 36(5): 800-803.

[4] 黄爱华, 王航, 唐卫东. 基于多阈值归一化分割的模糊图像边缘分割算法［J］. 半导体光电, 2017, 38(1): 142-145.

Huang Aihua, Wang Hang, Tang Weidong. Segmentation algorithms of fuzzy image edge based on multi-threshold normalized segmentation［J］. Semiconductor Optoelectronics, 2017, 38(1): 142-145.

[5] Long J, Shelhamer E, Darrell T. Fully convolutional networks for semantic segmentation［J］. IEEE Trans. Pattern Anal. Mach. Intell., 2017, 39(4): 640-651.

[6] Ronneberger O, Fischer P, Brox T. U-Net: Convolutional networks for biomedical image segmentation［C］// Inter. Conf. on Medical Image Computing and Computer-Assisted Intervention, 2015: 234-241.

[7] Noh H, Hong S, Han B. Learning deconvolution network for semantic segmentation［C］//Proc. of The IEEE Inter. Conf. on Computer Vision, 2015: 1520-1528.

[8] Lin T Y, Dollar P, Girshick R, et al. Feature pyramid networks for object detection［C］// IEEE Conf. on Computer Vision and Pattern Recognition, 2017: 2117-2125.

[9] He K, Zhang X, Ren S, et al. Deep residual learning for image recognition［C］// IEEE Conf. on Computer Vision and Pattern Recognition, 2016: 770-778.

[10] Chen L C, Papandreou G, Kokkinos I, et al. DeepLab: semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected CRFs［J］. IEEE Trans. Pattern Anal. Mach. Intell., 2016, 40(4): 834-848.

[11] Wang P, Chen P, Yuan Y, et al. Understanding convolution for semantic segmentation［C］// 2018 IEEE Winter Conf. on Applications of Computer Vision, 2018: 1451-1460.

李承珊, 蒋平, 崔雄文, 马震环, 雷涛. 基于编解码和局部增强的光电图像分割算法[J]. 半导体光电, 2018, 39(6): 892. LI Chengshan, JIANG Ping, CUI Xiongwen, MA Zhenhuan, LEI Tao. A Segmentation Algorithm of Optoelectric Images Based on Encoder-Decoder Structure and Local Image Enhancement[J]. Semiconductor Optoelectronics, 2018, 39(6): 892.

基于编解码和局部增强的光电图像分割算法

关于本站 Cookie 的使用提示

全站搜索

基于编解码和局部增强的光电图像分割算法

相关论文

相关资讯

关于本站 Cookie 的使用提示

全站搜索