towlbox

所属分类:其他
开发工具:Others
文件大小:2924KB
下载次数:1
上传日期:2017-08-15 14:07:55
上 传 者bohtdady
说明:  针对图象的二维自适应MDCT滤波工具箱,很实用 推荐给大家
(In view of the image two-dimensional adaptive toolbox, MDCT filtering is very practical Recommend to everyone)

文件列表:
Legal_Notice.txt (2929, 2006-03-08)
function_SADCT_RI_thresholding.dll (44032, 2006-04-12)
function_SADCT_RW_wiener.dll (114688, 2006-10-04)
function_SADCT_thresholding_fast.dll (159744, 2006-04-19)
function_SADCT_wiener_fast.dll (180224, 2006-04-19)
jpeg_read.dll (65536, 2003-10-30)
demo_SADCT_deblocking.m (47348, 2006-10-04)
demo_SADCT_deblurring.m (34838, 2006-10-04)
demo_SADCT_denoising.m (28843, 2006-10-04)
demo_SADCT_inverse_halftoning.m (31037, 2006-10-04)
function_LumChrom2rgb.m (2292, 2006-10-04)
function_rgb2LumChrom.m (3407, 2006-10-04)
function_shape_explorer.m (15319, 2006-04-06)
odemo_SADCT_color_denoising.m (33790, 2006-10-04)
function_AnisLPAICI8.p (19295, 2006-03-26)
image_Baboon512rgb.png (625179, 2005-10-29)
image_Barbara512.png (178047, 2005-09-16)
image_Barbara512v2.png (175255, 2005-09-16)
image_F16_512rgb.png (429132, 2005-12-08)
image_GreenPeppers512.png (165914, 2005-12-27)
image_House256.png (34399, 2005-09-16)
image_House256rgb.png (110937, 2005-10-29)
image_Lena512rgb.png (476195, 2005-10-29)
image_Peppers256.png (40181, 2005-10-05)
image_Peppers512rgb.png (507563, 2005-10-29)
matlab_temporary_jpeg_SA_DCT.jpg (2981, 2006-10-04)
JPEG_Toolbox_MATLAB_1.4.zip (64765, 2005-12-14)

Pointwise Shape-Adaptive DCT Demobox - Public release v1.40 (5 october 2006) ------------------------------------------------------------------------------------------------------------ (c) 2005-2006 Alessandro Foi Kostadin Dabov Tampere University of Technology, Tampere, FINLAND http://www.cs.tut.fi/~foi/SA-DCT ------------------------------------------------------------------------------------------------------------ The package comprises five demos demo_SADCT_denoising.m (denoising of grayscale images), demo_SADCT_color_denoising.m (denoising of color images), demo_SADCT_deblocking.m (deblocking and deringing of grayscale and color B-DCT compressed images), demo_SADCT_deblurring.m (deblurring of grayscale images), demo_SADCT_inverse_halftoning.m (reconstruction of grayscale images from binary halftones) and a function which shows the adaptive-shape neighborhoods which are used in the demos as supports of the Shape-Adaptive DCT transform function_shape_explorer.m (visualization of the adaptive-shape neighborhoods). The demos are open-source, and may be modified and tuned to be exploited with other data. Each demo calls a pair of functions in order to perform hard-thresholding and Wiener filtering in SA-DCT domain. In particular, the denoising and deblocking demos use function_SADCT_thresholding_fast.dll (hard-thresholding in SA-DCT domain), function_SADCT_wiener_fast.dll (Wiener-filtering in SA-DCT domain), while the deblurring demo uses the pair function_SADCT_RI_thresholding.dll (hard-thresholding in SA-DCT domain for regularized inverse), function_SADCT_RW_wiener.dll (Wiener-filtering in SA-DCT domain for regularized-Wiener inverse). Two more functions are called by the color denoising demo in order to perform color-space transformations: function_rgb2LumChrom.m (inverse transformation), function_LumChrom2rgb.m (forward transformation). Starting from v1.10, a new function function_AnisLPAICI8.p is introduced to perform the Anisotropic LPA-ICI adaptive-scale selection. It provides four to five times faster processing for the LPA-ICI part and requires much less memory than the previous general code based on conv2 and function_ICI code. Requirements: This demobox is developed for Matlab version 7 and the dll's are compiled for x86 (Pentium3, Athlon, or better)+Windows. This demobox requires a few functions from the Anisotropic Nonparametric Image Restoration Demobox of LASIP Project (available for download at http://www.cs.tut.fi/~lasip/). It is suggested that the Pointwise Shape-Adaptive DCT Demobox and the Anisotropic Nonparametric Image Restoration Demobox are installed in the same folder. The deblocking demo demo_SADCT_deblocking.m uses the function jpeg_read.dll from the Matlab JPEG Toolbox by Phil Sallee. A complete release of this toolbox is included in our distribution in the file JPEG_Toolbox_MATLAB_1.4.zip. What's new in this release: changes from version 1.40 to 1.31: + added: demo_SADCT_inverse_halftoning.m (demo for reconstruction of grayscale images from binary halftones) * updated: function_rgb2LumChrom.m (cleaner code) * updated: function_LumChrom2rgb.m (cleaner code) * updated: demo_SADCT_deblurring.m (cleaner code) * updated: demo_SADCT_deblocking.m (added option to disable coefficient alignment) * updated: demo_SADCT_denoising.m (added option to disable coefficient alignment) * updated: demo_SADCT_color_denoising.m (added option to disable coefficient alignment) * updated: function_SADCT_RW_wiener.dll (minor fix to regularization term) changes from version 1.31 to 1.30: ! bugfix: demo_SADCT_deblurring.m (fixed: calculation of RW variance) changes from version 1.30 to 1.20: * updated: demo_SADCT_deblocking.m (added option to disable coefficient alignment) * updated: demo_SADCT_denoising.m (added option to disable coefficient alignment) * updated: demo_SADCT_color_denoising.m (added option to disable coefficient alignment) * updated: function_SADCT_thresholding_fast.dll (now works with Pentium3/Athlon CPU, coeff. alignment can be disabled) * updated: function_SADCT_wiener_fast.dll (now works with Pentium3/Athlon CPU, coeff. alignment can be disabled) * updated: function_SADCT_RI_thresholding.dll (now works with Pentium3/Athlon CPU) * updated: function_SADCT_RW_wiener.dll (now works with Pentium3/Athlon CPU) changes from version 1.10 to 1.20: + added: demo_SADCT_deblocking.m (demo for deblocking and deringing of grayscale and color images) * updated: demo_SADCT_denoising.m (added blocksize parameter to constrain the maximum size of the filtering block) * updated: demo_SADCT_color_denoising.m (added blocksize parameter to constrain the maximum size of the filtering block) * updated: function_shape_explorer.m (added blocksize parameter to constrain the maximum size of the filtering block) changes from version 1.04 to 1.10: + added: function_AnisLPAICI8.p (enables much faster calculation of h_opt_Q, lower memory requirements) * updated: demo_SADCT_denoising.m (uses function_AnisLPAICI8, added speed-up factor parameter) * updated: demo_SADCT_color_denoising.m (uses function_AnisLPAICI8, added speed-up factor parameter) changes from version 1.03 to 1.04: ! bugfix: demo_SADCT_deblurring.m (fixed: boundary handling for calculation of variance of LPA estimates) changes from version 1.02 to 1.03: * updated: demo_SADCT_denoising.m (improved comments, cleaner code, new image cropping, and more visualization options) * updated: demo_SADCT_color_denoising.m (improved comments, cleaner code, new image cropping, and more visualization options) * updated: function_rgb2LumChrom.m (improved comments and cleaner code) * updated: function_LumChrom2rgb.m (improved comments and cleaner code) * updated: function_shape_explorer.m (improved comments) * updated: function_SADCT_thresholding_fast.dll (faster processing of constant blocks) changes from version 1.01 to 1.02: ! bugfix: function_SADCT_thresholding_fast.dll (fixed: number of harmonics for square blocks) changes from version 1.00 to 1.01: ! bugfix: function_SADCT_thresholding_fast.dll (fixed: preservation of mean for square blocks) + added: function_shape_explorer.m (shows adaptive-shape supports of SA-DCT) Future additions: A Linux version of the SA-DCT functions might be released in the future. Updates, examples, publications, presentations, etc. can be found at http://www.cs.tut.fi/~foi/SA-DCT Disclaimer: Any unauthorized use of these routines for industrial or profit-oriented activities is expressively prohibited. By downloading and/or using any of these files, you implicitly agree to all the terms of the TUT limited license (included in the file Legal_Notice.txt). Feedback: The authors of the Pointwise Shape-Adaptive DCT Demobox can be contacted at their respective email addresses firstname.lastname@tut.fi

近期下载者

相关文件


收藏者