首页>
外国专利>
SINGLE-IMAGE SUPER RESOLUTION AND DENOISING USING MULTIPLE WAVELET DOMAIN SPARSITY
SINGLE-IMAGE SUPER RESOLUTION AND DENOISING USING MULTIPLE WAVELET DOMAIN SPARSITY
展开▼
机译:单图像超分辨率和使用多小波域稀疏性进行降噪
展开▼
页面导航
摘要
著录项
相似文献
摘要
A computer-implemented method of enhancing images includes receiving one or more observed images, identifying wavelet bases, and determining a downsampling operator. A noise variance value is estimated and used to select a tuning parameter. A blurring kernel is estimated based on one or more system calibration parameter and used to determine a low-pass blurring filter operator. A cost function is created which generates one or more denoised super-resolution images based on the observed images and the plurality of wavelet bases. The cost function may include, for example, a sparsity inducing norm applied to the plurality of wavelet bases (with the tuning parameter applied to the sparsity inducing norm) and a constraint requiring the one or more denoised super-resolution images to be equal to a result of applying the low-pass blurring filter operator and the downsampling operator to the one or more denoised super-resolution images. The one or more denoised super-resolution images are generated by minimizing this cost function.
展开▼