基于机器学习的低分辨率图像增强和细节匹配方法
|
摘要
当前低分辨率图像增强和细节匹配方法具有细节易丢失、边缘模糊、无法适应图像平移、旋转等变化的弊端,导致图像增强与细节匹配性能低下。为此,提出一种新的基于机器学习的低分辨率图像增强和细节匹配方法。通过建立一个间隔最大的超平面获取最小二乘支持向量机分类器。在待处理低分辨率图像中选择一块图像,将图像的每个3×3邻域像素看作一个训练样本,通过最小二乘支持向量机法对其进行训练,输出增强像素点。通过复数小波对图像特征进行描述,利用最小二乘支持向量机获取最优判定准则函数,输出最优匹配的目标子图像。实验结果表明,所提方法有很高的峰值信噪比、边缘保持指数和等效视数,很低的归一化均方误差、均值和方差,整体性能优。
At present,the low resolution image enhancement and detail matching method is easy to be lost,blurred edge,can not adapt to the change of image translation,rotation and other defects,resulting in image enhancement and the performance of low matching performance. To this end, a new low resolution image enhancement and detail matching method based on machine learning is proposed. In the low resolution image to be processed,select an image,and the image of each 3 × 3 pixels as a training sample,the least squares support vector machine method to train the image,the output enhancement pixel. The image features are described by complex wavelet,and the least square support vector machine is used to obtain the optimal decision criterion function. The experimental results show that the proposed method has a high peak signal to noise ratio,the edge preserving index and the equivalent apparent number,and the low normalized mean square error,mean and variance,and the overall performance of the proposed method are excellent.
At present,the low resolution image enhancement and detail matching method is easy to be lost,blurred edge,can not adapt to the change of image translation,rotation and other defects,resulting in image enhancement and the performance of low matching performance. To this end, a new low resolution image enhancement and detail matching method based on machine learning is proposed. In the low resolution image to be processed,select an image,and the image of each 3 × 3 pixels as a training sample,the least squares support vector machine method to train the image,the output enhancement pixel. The image features are described by complex wavelet,and the least square support vector machine is used to obtain the optimal decision criterion function. The experimental results show that the proposed method has a high peak signal to noise ratio,the edge preserving index and the equivalent apparent number,and the low normalized mean square error,mean and variance,and the overall performance of the proposed method are excellent.
引文
1李可,杨奕臻,颜波.基于学习的图像智能适配显示技术.北京航空航天大学学报,2015;41(6):1147—1154Li Ke,Yang Yizhen,Yan Bo.Learning based intelligent image retargeting technique.Journal of Beijing University of Aeronautics and Astronautics,2015;41(6):1147—1154
2 盛帅,曹丽萍,黄增喜,等.基于改进稀疏编码的图像超分辨率算法.计算机应用,2014;34(2):562—566Sheng Shuai,Cao Liping,Huang Zengxi,et al.Image super-resolution algorithm based on improved sparse coding.Journal of Computer Applications,2014;34(2):562—566
3 李雪晴.基于模式对应与机器学习的HEVC降分辨率转码算法.电视技术,2016;40(1):1—6Li Xueqing.HEVC reduced resolution transcoding algorithm based on pattern correspondence and machine learning.TV Technology,2016;40 (1):1—6
4 黄铁军,郑锦,李波,等.多媒体技术研究:2013——面向智能视频监控的视觉感知与处理.中国图象图形学报,2014;19(11):1539—1562Huang Tiejun,Zheng Jin,Li Bo,et al.Visual perception and processing for intelligent video surveillance:a review.Journal of Image and Graphics,2014;19(11):1539—1562
5 尹义龙,杨公平,杨璐.指静脉识别研究综述.数据采集与处理,2015;30(5):933—939Yin Yilong,Yang Gongping,Yang Lu.Survey of finger vein recognition study.Journal of Data Acquisition&Processing,2015;30(5):933—939
6 Menten M J,Fast M F,Nill S,et al.WE-E-18A-02:Enhancement of lung tumor visibility by dual-energy X-ray imaging in an anthropomorphic chest phantom study.Medical Physics,2014;41(6):510
7 Khmag A,Abd R,Ramli S A R,et al.Image denoising algorithm using second generation wavelet transformation and principle component analysis.Research Journal of Applied Sciences Engineering&Technology,2014;8(3):367—377
8 肖进胜,高威,彭红,等.基于局部自相似性和奇异值分解的超采样图像细节增强.计算机学报,2016;39(7):1393—1406Xiao Jinsheng,Gao Wei,Peng Hong,et al.Detail enhancement for image super-resolution algorithm based on SVD and local selfsimilarity.Journal of Computer Science,2016;39(7):1393—1406
9 郑传远,李良福,肖樟树,等.一种改进的各向异性扩散深度图像增强算法.计算机工程与科学,2016;38(9):1823—1829Zheng Chuanyuan,Li Liangfu,Xiao Zhangshu,et al.A depth image enhancement algorithm based on improved anisotropic diffusion.Computer Engineering and Science,2016;38(9):1823—1829
10 刘艳莉,桂志国.基于形态学的可变权值匹配自适应图像增强算法.电子与信息学报,2014;36(6):1285—1291Liu Yanli,Gui Zhiguo.Adaptive image enhancement algorithm with variable weighted matching based on morphology.Journal of Electronics&Information Technology,2014;36(6):1285—1291
11 Liu Y L,Gui Z G.Adaptive image enhancement algorithm with variable weighted matching based on morphology.Journal of Electronics&Information Technology,2014;36(6):1285—1291
12 Pang Y,Yu B,Zhang X.Enhancement of the low resolution image quality using randomly sampled data for multi-slice MR imaging.Quantitative Imaging in Medicine&Surgery,2014;4(2):136—144
13 赵仁涛,王友余,李华德,等.多粒子群优化算法的远红外图像对比度增强.计算机仿真,2014;31(1):361—364Zhao Rentao,Wang Youyu,Li Huade,et al.Far infrared image contrast enhancement based on multi-particle swarm optimization.Computer Simulation,2014;31(1):361—364
14 蒋妍,翟金成.基于小波包分层净化的模糊图像细节增强算法.科技通报,2015;31(12):96—98Jiang Yan,Zhai Jincheng.Details enhancement algorithm of fuzzy image based on wavelet packet layered purification.Bulletin of Science and Technology,2015;31(12):96—98
15 黄向东,邓莉洁.一种基于Retinex的矿井非均匀照度图像增强算法.科学技术与工程,2014;14(5):141—144Huang Xiangdong,Deng Lijie.Algorithm of mine non-uniform illumination image enhancement based on retinex.Science Technology and Engineering,2014;14(5):141—144
16 Kang A S,Hong H.Near-real-time stereo matching method using temporal and spatial propagation of reliable disparity.Optical Engineering,2014;53(6):167—174
17 Zhao F,Zhang B,Zhao J,et al.The aerial image enhancement algorithm based on histogram specification induced by the adaptive function.Journal of Optoelectronics Laser,2015;26(1):192—198
18 万荣泽,莫洪武,余思东.基于超混沌DNA计算优化OTP算法的文本图像加密算法.计算机测量与控制,2014;22(10):3278—3281Wan Rongze,Mo Hongwu,Yu Sidong.Document and image encryption based on OTP optimized by hyper-chaos mapping DNA computing.Computer Measurement&Control,2014;22(10):3278—3281
19 戴乐,王凯凯,朱烨.基于多重分形和改进BP算法的环境图像识别.电子设计工程,2016;24(17):167—170Dai Le,Wang Kaikai,Zhu Ye.Environment images based on multifractal and improved BP algorithm recognition.Electronic Design Engineering,2016;24(17):167—170
20 Anitha A,Kumar B R.A new image enhancement algorithm using just noticeable difference and histogram equalization.International Journal of Engineering Trends&Technology,2014;9(6):282—287
2 盛帅,曹丽萍,黄增喜,等.基于改进稀疏编码的图像超分辨率算法.计算机应用,2014;34(2):562—566Sheng Shuai,Cao Liping,Huang Zengxi,et al.Image super-resolution algorithm based on improved sparse coding.Journal of Computer Applications,2014;34(2):562—566
3 李雪晴.基于模式对应与机器学习的HEVC降分辨率转码算法.电视技术,2016;40(1):1—6Li Xueqing.HEVC reduced resolution transcoding algorithm based on pattern correspondence and machine learning.TV Technology,2016;40 (1):1—6
4 黄铁军,郑锦,李波,等.多媒体技术研究:2013——面向智能视频监控的视觉感知与处理.中国图象图形学报,2014;19(11):1539—1562Huang Tiejun,Zheng Jin,Li Bo,et al.Visual perception and processing for intelligent video surveillance:a review.Journal of Image and Graphics,2014;19(11):1539—1562
5 尹义龙,杨公平,杨璐.指静脉识别研究综述.数据采集与处理,2015;30(5):933—939Yin Yilong,Yang Gongping,Yang Lu.Survey of finger vein recognition study.Journal of Data Acquisition&Processing,2015;30(5):933—939
6 Menten M J,Fast M F,Nill S,et al.WE-E-18A-02:Enhancement of lung tumor visibility by dual-energy X-ray imaging in an anthropomorphic chest phantom study.Medical Physics,2014;41(6):510
7 Khmag A,Abd R,Ramli S A R,et al.Image denoising algorithm using second generation wavelet transformation and principle component analysis.Research Journal of Applied Sciences Engineering&Technology,2014;8(3):367—377
8 肖进胜,高威,彭红,等.基于局部自相似性和奇异值分解的超采样图像细节增强.计算机学报,2016;39(7):1393—1406Xiao Jinsheng,Gao Wei,Peng Hong,et al.Detail enhancement for image super-resolution algorithm based on SVD and local selfsimilarity.Journal of Computer Science,2016;39(7):1393—1406
9 郑传远,李良福,肖樟树,等.一种改进的各向异性扩散深度图像增强算法.计算机工程与科学,2016;38(9):1823—1829Zheng Chuanyuan,Li Liangfu,Xiao Zhangshu,et al.A depth image enhancement algorithm based on improved anisotropic diffusion.Computer Engineering and Science,2016;38(9):1823—1829
10 刘艳莉,桂志国.基于形态学的可变权值匹配自适应图像增强算法.电子与信息学报,2014;36(6):1285—1291Liu Yanli,Gui Zhiguo.Adaptive image enhancement algorithm with variable weighted matching based on morphology.Journal of Electronics&Information Technology,2014;36(6):1285—1291
11 Liu Y L,Gui Z G.Adaptive image enhancement algorithm with variable weighted matching based on morphology.Journal of Electronics&Information Technology,2014;36(6):1285—1291
12 Pang Y,Yu B,Zhang X.Enhancement of the low resolution image quality using randomly sampled data for multi-slice MR imaging.Quantitative Imaging in Medicine&Surgery,2014;4(2):136—144
13 赵仁涛,王友余,李华德,等.多粒子群优化算法的远红外图像对比度增强.计算机仿真,2014;31(1):361—364Zhao Rentao,Wang Youyu,Li Huade,et al.Far infrared image contrast enhancement based on multi-particle swarm optimization.Computer Simulation,2014;31(1):361—364
14 蒋妍,翟金成.基于小波包分层净化的模糊图像细节增强算法.科技通报,2015;31(12):96—98Jiang Yan,Zhai Jincheng.Details enhancement algorithm of fuzzy image based on wavelet packet layered purification.Bulletin of Science and Technology,2015;31(12):96—98
15 黄向东,邓莉洁.一种基于Retinex的矿井非均匀照度图像增强算法.科学技术与工程,2014;14(5):141—144Huang Xiangdong,Deng Lijie.Algorithm of mine non-uniform illumination image enhancement based on retinex.Science Technology and Engineering,2014;14(5):141—144
16 Kang A S,Hong H.Near-real-time stereo matching method using temporal and spatial propagation of reliable disparity.Optical Engineering,2014;53(6):167—174
17 Zhao F,Zhang B,Zhao J,et al.The aerial image enhancement algorithm based on histogram specification induced by the adaptive function.Journal of Optoelectronics Laser,2015;26(1):192—198
18 万荣泽,莫洪武,余思东.基于超混沌DNA计算优化OTP算法的文本图像加密算法.计算机测量与控制,2014;22(10):3278—3281Wan Rongze,Mo Hongwu,Yu Sidong.Document and image encryption based on OTP optimized by hyper-chaos mapping DNA computing.Computer Measurement&Control,2014;22(10):3278—3281
19 戴乐,王凯凯,朱烨.基于多重分形和改进BP算法的环境图像识别.电子设计工程,2016;24(17):167—170Dai Le,Wang Kaikai,Zhu Ye.Environment images based on multifractal and improved BP algorithm recognition.Electronic Design Engineering,2016;24(17):167—170
20 Anitha A,Kumar B R.A new image enhancement algorithm using just noticeable difference and histogram equalization.International Journal of Engineering Trends&Technology,2014;9(6):282—287