基于深度卷积生成对抗网络的图像识别算法

刘恋秋

doi:doi:10.3788/yjyxs20203504.0383

液晶与显示, 2020, 35 (4): 383, 网络出版: 2020-05-30

基于深度卷积生成对抗网络的图像识别算法

Image recognition algorithms based on deep convolution generative adversarial network

论文大纲

刘恋秋 ^*

作者单位

重庆财经职业学院, 重庆 402160

生成对抗网络深度卷积特征提取计算机视觉 generate confrontation network deep convolution feature extraction computer vision

摘要

针对传统深度卷积生成网络收敛速度慢、稳定性较差的问题, 本文在传统深度卷积生成对抗网络的基础上, 提出了深度卷积生成对抗网络的优化算法。首先在预处理部分, 融合了Canny算子和Prewitt算子的多个方向的卷积核来初始化输入图片参数,同时训练模块。为了减少训练时间, 将训练分为3个阶段, 每个阶段都采用不同的损失函数, 从而提升网络的收敛速度及识别效果。最后再将训练后的判别网络中的卷积神经网络用来提取图像特征。LFW和CIFAR-100的实验证明, 本文提出的算法具有很高的可行性和有效性, 比传统生成对抗网络、CNN等图像识别具有更高的识别成功率, 达到89.5%, 为生成对抗网络在计算机视觉领域的应用提供了有益的参考。

Abstract

In view of the slow convergence and poor stability of traditional deep convolutional generation networks, this paper combines the Connaught kernels of multiple directions of Canny operator and Prewitt operator on the basis of traditional deep convolution generation confrontation network. To initialize the input picture parameters and reduce the training time, the training is divided into three stages, each stage adopts a different loss function, thereby improving the convergence speed and recognition effect of the network and then discriminating the volume in the network after training. The neural network is used to extract image features. The experiments of LFW and CIFAR-100 prove that the proposed algorithm is highly feasible and effective, and provides a useful reference for generating anti-network applications in the field of image recognition.

参考文献

[1] GOODFELLOW I J, POUGET-ABADIE J, MIRZA M, et al. Generative adversarial networks ［J］. Advances in Neural Information Processing Systems, 2014, 3: 2672-2680.

[2] RADFORD A, METZ L, CHINTALA S. Unsupervised representation learning with deep convolutional generative adversarial networks ［J］. Computer Science, 2015, 12(36): 1280-1287.

[3] SALIMANS T, GOODFELLOW I, ZAREMBA W, et al. Improved techniques for training GANs ［J］. ArXiv, 2016, 606.03498v1 .

[4] 丁鹏, 张叶, 刘让,等. 结合形态学和Canny算法的红外弱小目标检测［J］. 液晶与显示, 2016, 31(8): 793-800.

DING P, ZHANG Y, LIU R, et al. Infrared weak target detection based on morphology and Canny algorithm ［J］. Chinese Journal of Liquid Crystal and Displays, 2016, 31(8): 793-800. (in Chinese)

[5] KIM Y. Convolutional neural networks for sentence classification ［J］. ArXiv, 2014,1408.5882.

[6] 金志刚李静昆. 基于对象性和多层线性模型的协同显著性检测［J］. 光学精密工程, 2019, 27(8): 1845-1853.

JIN Z G, LI J K. Co-saliency detection based onobjectness and multi-layer linear model ［J］. Optics and Precision Engineering, 2019, 27(8): 1845-1853. (in Chinese)

[7] 潘仙张张石清郭文平. 多模深度卷积神经网络应用于视频表情识别［J］. 光学精密工程, 2019, 27(4): 963-970.

PAN X Z, ZHANG S Q, GUO W P. Video-based facial expression recognition using multimodal deepconvolutional neural networks ［J］. Optics and Precision Engineering, 2019, 27(4): 963-970. (in Chinese)

[8] 张坤华, 谭志恒, 李斌. 结合粒子群优化和综合评价的脉冲耦合神经网络图像自动分割［J］. 光学精密工程, 2018, 26(4): 962-970.

ZHANG K H, TAN Z H, LI B. Automated image segmentation based on pulse coupled neural network withpartide swarm optimization and comprehensive evaluation ［J］. Optics and Precision Engineering, 2018, 26(4): 962-970. (in Chinese)

[9] 吕永标, 赵建伟, 曹飞龙. 基于复合卷积神经网络的图像去噪算法［J］. 模式识别与人工智能, 2017, 30(2): 97-105.

LYU Y B, ZHAO J W, CAO F L. Image denoising algorithm based on compound convolution neural network ［J］. Pattern Recognition and Artificial Intelligence, 2017, 30 (2): 97-105. (in Chinese)

[10] 王秀席, 王茂宁, 张建伟,等. 基于改进的卷积神经网络lenet-5的车型识别方法［J］. 计算机应用研究, 35(7); 301-304.

WANG X X, WANG M N, ZHANG J W, et al. Vehicle recognition method based on improved convolutional neural network lenet-5 ［J］. Computer Application Research, 35 (7): 301-304. (in Chinese)

[11] 许赟杰, 徐菲菲. 基于arcrelu函数的神经网络激活函数优化研究［J］. 数据采集与处理, 2019, 34(3): 517-529.

XU Y J, XU F F. Optimization of activation function in neural network based on arcrelu function ［J］. Data Acquisition and Processing, 2019, 34(3): 517-529. (in Chinese)

[12] ZHENG Z, ZHENG L, YANG Y. Unlabeled samples generated by GAN improve the person re-identification baseline in vitro ［C］. IEEE International Conference on Computer Vision, USA: Honolulu, IEEE, 2017.

[13] ABADI M, AGARWAL A, BARHAM P, et al. TensorFlow: large-scale machine learning on heterogeneous distributed systems ［J］. ArXiv, 2016,1603.04467.

[14] GOODFELLOW I J, WARDE-FARLEY D, MIRZA M, et al. Maxout networks ［J］. Computer Science, 2013: 1319-1327.

[15] TAIGMAN Y, MING Y, RANZATO M, et al. DeepFace: closing the gap to human-level performance in face verification［C］. IEEE Conference on Computer Vision & Pattern Recognition, USA: Columbus, Ohio, IEEE, 2014.

刘恋秋. 基于深度卷积生成对抗网络的图像识别算法[J]. 液晶与显示, 2020, 35(4): 383. LIU Lian-qiu. Image recognition algorithms based on deep convolution generative adversarial network[J]. Chinese Journal of Liquid Crystals and Displays, 2020, 35(4): 383.

基于深度卷积生成对抗网络的图像识别算法

关于本站 Cookie 的使用提示

全站搜索

基于深度卷积生成对抗网络的图像识别算法

相关论文

相关资讯

关于本站 Cookie 的使用提示

全站搜索