updated demo

demo_script.m

@@ -60,9 +60,23 @@
 [C,f,P,S,YrA] = update_temporal_components(Yr,A,b,Cin,fin,P,options);
 
 %% classify components
+
 [ROIvars.rval_space,ROIvars.rval_time,ROIvars.max_pr,ROIvars.sizeA,keep] = classify_components(Y,A,C,b,f,YrA,options);
-A_keep = A(:,keep);
-C_keep = C(keep,:);
+
+%% further classification with cnn_classifier
+try  % matlab 2017b or later is needed
+    [ind,value] = cnn_classifier(A,[d1,d2],'cnn_model',0.2);
+catch
+    ind = true(size(A,2),1);
+end
+%% display kept and discarded components
+A_keep = A(:,(keep & ind));
+C_keep = C((keep & ind),:);
+figure;
+    subplot(121); montage(extract_patch(A(:,(keep & ind)),[d1,d2],[30,30]),'DisplayRange',[0,0.15]);
+        title('Kept Components');
+    subplot(122); montage(extract_patch(A(:,~(keep & ind)),[d1,d2],[30,30]),'DisplayRange',[0,0.15])
+        title('Discarded Components');
 %% merge found components
 [Am,Cm,K_m,merged_ROIs,Pm,Sm] = merge_components(Yr,A_keep,b,C_keep,f,P,S,options);
 

utilities/extract_patch.m

@@ -17,6 +17,7 @@
 nd = length(dims);
 if nd == 2; dims(3) = 1; patch_size(3) = 1; end
 K = size(A,2);
+A = A/spdiags(sqrt(sum(A.^2,1))'+eps,0,K,K);      % normalize to sum 1 for each compoennt
 cm = com(A,dims(1),dims(2),dims(3));
 xx = -ceil(patch_size(1)/2-1):floor(patch_size(1)/2);
 yy = -ceil(patch_size(2)/2-1):floor(patch_size(2)/2);