摘要
对互联网中虚假图像的识别,能够有效提高网络广告信息传播的真实度。对虚假广告数字图像的优化识别,需要定义图像真实的边缘,对边缘的像素级进行定位,完成虚假广告图像的识别。传统方法利用形态学优化的边缘检测理论检测出图像的边界,得到图像的边缘特征,但忽略了定位图像像素,导致识别精度偏低。提出基于边缘羽化检测的互联网虚假广告数字图像优化识别方法。利用双树复小波变换理论进行图像分解,得到分解后的图像纹理数据,进行数字图像类似纹理的再提取,得到图像纹理数据的门限值,重新构建数字图像纹理轮廓,引入局部清晰度来确定图像上的边缘点,检测图像的虚假边缘,对数字图像虚假边缘的像素级进行定位,以此为依据进行互联网虚假广告数字图像优化识别。实验证明,所提方法识别精度较高,为保护互联网下数字图像的真实性和完整性奠定了基础。
The identification of false image in Internet can effectively improve the authenticity of network advertising information dissemination. The optimization and recognition of false advertising digital images needs to define the true edge of image. The traditional method uses the edge detection theory based on morphological optimization to detect the image edge, which ignores to locate the pixel of image, resulting in low recognition accuracy. A method for optimizing and recognizing false advertising digital image on Internet based on edge feather detection is proposed. This method used dual - tree complex wavelet transform theory to decompose image, so that decomposed data of image texture was obtained. Then, the method extracted the similar texture of digital image again, so as to obtain threshold value of image texture data. After reconstructing digital image texture, we introduced the local definition to determine edge points on image. On the basis of positioning the pixel level of digital image false edges, the false advertising digital image on Internet was optimized and recognized. Simulations show that the proposed method has high recognition accuracy, which lays a foundation for protecting the authenticity and integrity of digital images on the internet.
作者
全蕾
QUAN Lei(College of Yangtze River, Donghua University of science and technology, Jiangxi, Fuzhou 344100, China)
出处
《计算机仿真》
北大核心
2018年第6期447-450,共4页
Computer Simulation
关键词
互联网
虚假广告
数字图像
优化识别
Internet
False advertisement
Digital image
Optimization recognition
