曲波变换和最小二乘支持向量机的图像压缩算法
发布时间：2019-08-05 10:57
【摘要】：为了提高图像压缩质量,针对传统压缩算法的不足,提出一种曲波变换和最小二乘支持向量机相融合的图像压缩算法。首先采用曲波变换把图像分解为不同尺度和不同方向的曲波系数,并采用熵编码对粗尺度层曲波系数进行压缩,然后利用最小二乘支持向量机对细尺度层中不同方向的曲波系数进行学习,并通过和声搜索算法优化最小二乘支持向量机,实现细尺度层曲波数的压缩,最后采用图像压缩仿真实验测试其性能。结果表明,曲波变换和最小二乘支持向量机相融合的图像压缩算法提高了图像压缩的峰值信噪比,加快了图像压缩的速度,获得了更好的图像压缩效果。
[Abstract]:In order to improve the quality of image compression, an image compression algorithm based on curved wave transform and least square support vector machine is proposed to overcome the shortcomings of traditional compression algorithms. Firstly, the curved wave transform is used to decompose the image into curved wave coefficients of different scales and directions, and entropy coding is used to compress the curved wave coefficients of rough scale layer. Then, the curved wave coefficients in different directions in fine scale layer are learned by least square support vector machine, and the least square support vector machine is optimized by harmony search algorithm to realize the compression of curved wave number of fine scale layer. Finally, the performance of the image compression simulation experiment is tested. The results show that the image compression algorithm based on curved wave transform and least square support vector machine improves the peak signal-to-noise ratio (PSNR) of image compression, accelerates the speed of image compression, and obtains better image compression effect.
【作者单位】： 黄淮学院信息工程学院;
【基金】：国家自然科学基金资助项目(编号:61372058) 辽宁省高等学校优秀人才支持计划项目(编号:LR2013012)
【分类号】：TN911.73
，
本文编号：2523086
[Abstract]:In order to improve the quality of image compression, an image compression algorithm based on curved wave transform and least square support vector machine is proposed to overcome the shortcomings of traditional compression algorithms. Firstly, the curved wave transform is used to decompose the image into curved wave coefficients of different scales and directions, and entropy coding is used to compress the curved wave coefficients of rough scale layer. Then, the curved wave coefficients in different directions in fine scale layer are learned by least square support vector machine, and the least square support vector machine is optimized by harmony search algorithm to realize the compression of curved wave number of fine scale layer. Finally, the performance of the image compression simulation experiment is tested. The results show that the image compression algorithm based on curved wave transform and least square support vector machine improves the peak signal-to-noise ratio (PSNR) of image compression, accelerates the speed of image compression, and obtains better image compression effect.
【作者单位】： 黄淮学院信息工程学院;
【基金】：国家自然科学基金资助项目(编号:61372058) 辽宁省高等学校优秀人才支持计划项目(编号:LR2013012)
【分类号】：TN911.73
，
本文编号：2523086
本文链接：https://www.wllwen.com/kejilunwen/wltx/2523086.html
