% Function written for NanoLocz-lib (2025) % FILTER_MOVIE Applies one or two filters to a 3D image sequence (movie). % % target = FILTER_MOVIE(target, filt1, strength1) % Applies the filter specified by filt1 with strength strength1 to the 3D % image stack 'target'. % % Result = FILTER_MOVIE(target, filt1, strength1, filt2, strength2) % Applies two filters sequentially: filt1 with strength1, then filt2 with strength2. % % INPUTS: % target - 3D image stack (x, y, frames) to be filtered. % filt1 - Name of the first filter to apply (string). % strength1 - Strength parameter for the first filter. % filt2 - (Optional) Name of the second filter to apply (string). % strength2 - (Optional) Strength parameter for the second filter. % % OUTPUT: % target - Filtered 3D image stack. % % Available Filters (case-sensitive names): % 'Gaussian' - Gaussian smoothing (strength = standard deviation) % '-Gaussian' - Subtract Gaussian-smoothed image (high-pass effect) % 'Sphere' - Sphere-shaped convolution smoothing % 'Wiener' - Adaptive Wiener filtering (strength = neighborhood size) % 'Non-local-means' - Non-local means denoising ('DegreeOfSmoothing') % 'Disk' - Disk-shaped convolution smoothing % 'ButterLP' - Butterworth low-pass filter (strength = cutoff frequency) % 'ButterHP' - Butterworth high-pass filter (strength = cutoff frequency) % 'ButterBP' - Butterworth band-pass filter ([low high] cutoff frequencies) % 'Laplacian' - Laplacian edge enhancement (strength = weight) % '-Average' - Subtract mean image over time (high-pass over time) % 'Deconv' - Deconvolution with sphere kernel (strength = size) % 'Peak sharp' - Peak sharpening along rows (strength = sharpening factor) % 'High-pass' - Frequency-domain high-pass filtering (strength = scale factor) % 'RAvg' - Running average over time (strength = window size) % % Notes: % - If no filters are desired, set strength to 0. % - Filters are applied in sequence if two are provided. % - The function modifies the input image stack in place. % - Peak sharp function target = filter_movie(target, filt1, strength1, filt2, strength2) filt{1} = filt1; strength{1} = strength1; if nargin >3 filt{2} = filt2; strength{2} = strength2; end n = size(target,3); for i =1:numel(filt) if strength{i}>0 if strcmp('Gaussian', filt{i}) if strength{i} ==0 strength{i} = 0.01; else end target = imgaussfilt(target,strength{i}); elseif strcmp('-Gaussian',filt{i}) if strength{i} ==0 strength{i} = 0.01; else end target = target - imgaussfilt(target,strength{i}); elseif strcmp('Wiener', filt{i}) if strength{i} >=2 for j=1:n target(:,:,j) = wiener2( target(:,:,j),[strength{i},strength{i}]); end end elseif strcmp('Non-local-means',filt{i}) if strength{i} ==0 strength{i} = 0.01; else end for j=1:n target(:,:,j) = imnlmfilt(target(:,:,j),'DegreeOfSmoothing',strength{i}); end elseif strcmp('Disk',filt{i}) P = strength{i}+1; offs = round(P)- P; [hx,hy] = ndgrid(-(P):(P)); hx = hx-offs; hy = hy-offs; H = real(sqrt((-(hx.^2)-(hy.^2)+P^2)))>P/2; H=H./(sum(H(:))); if offs>0 H = H(2:end , 2:end); end for j=1:n target(:,:,j) = imfilter(target(:,:,j),H); end elseif strcmp('Sphere',filt{i}) P = strength{i}+1; offs = round(P)- P; [hx,hy] = ndgrid(-(P):(P)); hx = hx-offs; hy = hy-offs; H = real(sqrt((-(hx.^2)-(hy.^2)+P^2))); H=H./(sum(H(:))); if offs>0 H = H(2:end , 2:end); end for j=1:n target(:,:,j) = imfilter(target(:,:,j),H); end elseif strcmp('ButterLP', filt{i}) fs = 1; % Sampling frequency in Hz order = 1; % Order of the filter try [b, a] = butter(order, strength{i}/(fs/2), 'low'); catch [b, a] = butter(order, 0.49/(fs/2), 'high'); end sz = size(target); for j=1:n img1 = target(:,:,j); for ii = 1:sz(1) img2(ii,:) = filtfilt(b, a, img1(ii,:)); end target(:,:,j) = img2; end elseif strcmp('ButterHP', filt{i}) fs = 1; % Sampling frequency in Hz order = 1; % Order of the filter try [b, a] = butter(order, strength{i}/(fs/2), 'high'); catch [b, a] = butter(order, 0.49/(fs/2), 'high'); end sz = size(target); for j=1:n img1 = target(:,:,j); for ii = 1:sz(1) img2(ii,:) = filtfilt(b, a, img1(ii,:)); end target(:,:,j) = img2; end elseif strcmp('Laplacian', filt{i}) h = fspecial('laplacian',0.2); %h = [-1 -1 -1; -1 8 -1; -1 -1 -1]; %target = imfilter(target, h, 'replicate'); lap = imfilter(imgaussfilt(target,0.6),h,'replicate'); lap = rescale(lap); target = rescale(-strength{i}*lap+target); elseif strcmp('-Average', filt{i}) d_av = mean(target,3); target = target - d_av.*strength{i}; elseif strcmp('Deconv', filt{i}) if strength{i}>0 P = strength{i}+1; offs = round(P)- P; [hx,hy] = ndgrid(-(P):(P)); hx = hx-offs; hy = hy-offs; H = real(sqrt((-(hx.^2)-(hy.^2)+P^2))); H=H./(sum(H(:))); if offs>0 H = H(2:end , 2:end); end target = deconvlucy(target,H,1); end elseif strcmp('Peak sharp', filt{i}) if strength{i}>0 sz = size(target); for j=1:n for ii = 1:sz(1) img2(ii,:,j) = sharpen(1:sz(2),target(ii,:,j),strength{i},0,3); end end target = img2; end elseif strcmp('High-pass', filt{i}) for j=1:n [mx, ny] =size( target(:,:,j)); ft = fft2( target(:,:,j)); f_shift = fftshift(ft); ps = mx/2; qs = ny/2; for ii=1:mx for jj=1:ny distance = sqrt((ii-ps)^2+(jj-qs)^2); low_filter(ii,jj) = 1-exp(-(distance)^2/(2*(strength{i}))); end end filter_apply = f_shift.*low_filter; image_orignal = ifftshift(filter_apply); target(:,:,j) = abs(ifft2(image_orignal)); end elseif strcmp('RAvg', filt{i}) target = movmean(target,strength{i},3); elseif strcmp('ButterBP', filt{i}) fs = 1; % Sampling frequency in Hz order = 1; % Order of the filter low_cutoff = strength{i}(1); % Lower cutoff frequency high_cutoff = strength{i}(2); % Upper cutoff frequency try [b, a] = butter(order, [low_cutoff high_cutoff]/(fs/2), 'bandpass'); catch [b, a] = butter(order, [0.01 0.49], 'bandpass'); % Fallback values end sz = size(target); for j = 1:n img1 = target(:,:,j); for ii = 1:sz(1) img2(ii,:) = filtfilt(b, a, img1(ii,:)); end target(:,:,j) = img2; end end end end