86120636.rar_图形/文字识别

function y = nldifc( u, lambda, sigma, m, stepsize, nosteps, varargin) % % NLDIFC Color Nonlinear Diffusion % % y = NLDIFC( u, lambda, sigma, m, stepsize, steps, verbose, drawstep ) returns the % image y as the result of the application of the Non Linear diffusion % to the image u (See NLDIF for details). % % NLDIFC performs an independent diffusion in each color channel of u. If the used % parameters are scalars or row vectors, the same parameter will be used for every % channel. To use different parameters for each channel, each parameter must be passed % as a 1xCH cell array. String parameters cannot be defined for each channel individually % % Example: y = nldif(u, {3 4 5}, 1, 10, .2, 10) will diffuse ch 1 using lambda = 3 % ch 2 with lambda = 4 and ch3 with lambda = 5. Every other parameter will be the % same for the three channels. % % y = NLDIFC( ..., 'alt1') computes the diffusivity based on the p-norm of the gradient % in each channel and filters all channels at once instead of using each channel separately. % By default p = 2. For other norms use : y = NLDIFC(..., 'alt1','norm',p). p can be any % number or 'inf'. % % See also: NLDIF, PMDIF, EEDIF, CEDIF, DIFFUSIVITY, GSDPLOT, FLUXPLOT % Nch = size(u,3); [verbose, drawstep, imscale, plotgrad, plotflux, dfstep, aos, alt1, pnorm] = parse_inputs(varargin{:}); % Check Inputs [lambda, sigma, m, stepsize, nosteps, verbose, drawstep, dfstep] = check_inputs(Nch,lambda, sigma, m, stepsize, nosteps, verbose, drawstep, dfstep, alt1); if alt1 == 0 y = zeros(size(u)); for ch = 1 : size(u,3) % MAKEVAR v = makevar(verbose{ch},drawstep{ch},imscale,plotgrad,plotflux,dfstep{ch},aos,ch); disp( ['Processing channel : ', int2str(ch) , ' / ' , int2str(Nch) ] ); eval([ 'y(:,:,' , int2str(ch) , ') = nldif(u(:,:, ' , int2str(ch) , ' ),lambda{' , int2str(ch) ,'},sigma{' , int2str(ch) ,'},m{' , int2str(ch) ,'},stepsize{' , int2str(ch) ,'},nosteps{' , int2str(ch) ,'}, ' , v , ');' ]) ; end y = scale(y,[0 1]); % Scaling %Plotting if any( cat(1,verbose{:}) ) subplot(1,2,1) imagesc(u); title('Original Image') subplot(1,2,2) imagesc(y) title('Non Linear Diffusion') end else % alt1 = 1 :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: % Variable initialization dif_time = 0; if strcmp(class(u),'double') y = u; else y = double(u); end g = cell(1,Nch); for j = 1 : Nch g{j} = zeros( size(y(:,:,1)) ); end % Verifying inputs if alt1 == 0 for j = 1 : Nch [lambda{j}, sigma{j}, stepsize{j}] = verify_inputs(lambda{j}, sigma{j}, stepsize{j}, nosteps{j}); end else for j = 1 : Nch [lambda{j}, sigma{j}, stepsize] = verify_inputs(lambda{j}, sigma{j}, stepsize, nosteps); end end % Initial Drawing if verbose figure(verbose); subplot(1,2,1); if ndims(y) == 3 image(scale(y,[0 1])); else if strcmp(imscale,'imscale') imagesc(y); else image(y); end colorbar end title('Original Image'); drawnow; difplot(y, 0, 0, 'Non Linear Diffusion', verbose, imscale); end % Calculate Cm constant for j = 1 : Nch Cm{j}= Cmcalc(m{j}); end for i=1:nosteps if mod(i-1,dfstep) == 0 % diffusivity recalc step grad = zeros( size(y(:,:,1)) ); % Calculate gradient of smoothed image for j = 1 : Nch [gradx, grady] = gsderiv(y(:,:,j), sigma{j}(i), 1); if strcmp(pnorm,'inf') grad = max(grad, sqrt(gradx.^2 + grady.^2) ); elseif pnorm == 2 grad = grad + gradx.^2 + grady.^2; elseif pnorm == 1 grad = grad + sqrt(gradx.^2 + grady.^2); else grad = grad + (sqrt(gradx.^2 + grady.^2)).^pnorm; end end if ~strcmp(pnorm,'inf') & pnorm~=1 grad = grad.^(1/pnorm); % Not necessary to do anything if pnorm == 'inf' or 1. end if plotgrad figure(verbose+1) imagesc(grad); title('Grad. Magnitude') colorbar axis off end % Calculate difusivity if samedif(lambda,m,i) % If lambda and m are the same for all channels in this step g{1} = 1 - exp(-Cm{1}./( (grad+eps)./lambda{1}(i)).^m{1}); for j = 2 : Nch g{j} = g{1}; end else for j = 1 : Nch g{j} = 1 - exp(-Cm{j}./( (grad+eps)./lambda{j}(i)).^m{j}); % grad + eps to avoid division by zero (eps -> 0) end end end % diff. recalc % Calculate dy/dt if aos if plotflux yo = y; end for j = 1 : Nch y(:,:,j) = aosiso(y(:,:,j),g{j},stepsize(i)); % updating end else for i = 1 : Nch dy(:,:,1) = isodifstep(y(:,:,1), g{j}); end y = y + stepsize(i) * dy; % updating end % Calculate diffusion time dif_time = dif_time + stepsize(i); % Plot actualization if verbose & ~mod(i,drawstep) difplot(y, dif_time, i, 'Non Linear Diffusion', verbose, imscale); if plotflux figure(verbose+plotgrad+1) if aos fluxplot(g,grad, norm3(abs(y-yo)), 'axisoff') else fluxplot(g,grad, norm3( abs(stepsize(i).*(dy)) ), 'axisoff') end end end % Plot actualization end % i = 1 : nosteps % Last plot difplot(y, dif_time, i, 'Non Linear Diffusion', verbose, imscale); if plotflux figure(verbose+plotgrad+1) if aos fluxplot(g,grad, norm3(abs(y-yo)), 'axisoff') else fluxplot(g,grad, norm3( abs(stepsize(i).*(dy)) ), 'axisoff') end end end % alt1 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% function y = cellcomp(x,n) for i = 1 : n y{i} = x; end %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% function [verbose, drawstep, imscale, plotgrad, plotflux, dfstep, aos, alt1, pnorm] = parse_inputs(varargin) pnorm = 2; pnormpos = -1; alt1 = 0; aos = 0; dfsteppos = -1; dfstep = 1; verbose = -1; drawstep = -1; imscale = 'null'; plotgrad = 0; plotflux = 0; for i = 1 : length(varargin) flag = 0; if i == dfsteppos flag = 1; end if i == pnormpos flag = 1; end if strcmp(varargin{i},'hsv') hsv = 1; flag = 1; elseif strcmp(varargin{i},'norm') pnorm = varargin{i+1}; flag = 1; pnormpos = i+1; elseif strcmp(varargin{i},'alt1') alt1 = 1; flag = 1; elseif strcmp(varargin{i},'imscale') imscale = 'imscale'; flag = 1; elseif strcmp(varargin{i},'grad') plotgrad = 1; flag = 1; elseif strcmp(varargin{i},'flux') plotflux = 1; flag = 1; elseif strcmp(varargin{i},'dfstep') dfstep = varargin{i+1}; flag = 1; dfsteppos = i+1; elseif strcmp(varargin{i},'aos') aos = 1; flag = 1; end if flag == 0 & verbose == -1 verbose = varargin{i}; flag = 1; end if flag == 0 & drawstep == -1 drawstep = varargin{i}; flag = 1; end if flag == 0 error('Too many parameters !') return end end if verbose == -1 verbose = 0; end