def init_rois(self, n_components=100, show=False):
Ain, Cin, center = greedyROI2d(self.Y,
nr=n_components,
gSig=[2, 2],
gSiz=[7, 7],
use_median=False)
Cn = np.mean(self.Y, axis=-1)
if show:
pl1 = pl.imshow(Cn, interpolation='none')
pl.colorbar()
pl.scatter(x=center[:, 1], y=center[:, 0], c='m', s=40)
pl.axis((-0.5, self.Y.shape[1] - 0.5, -0.5, self.Y.shape[0] - 0.5))
pl.gca().invert_yaxis()
active_pixels = np.squeeze(np.nonzero(np.sum(Ain, axis=1)))
Yr = np.reshape(self.Y,
(self.Y.shape[0] * self.Y.shape[1], self.Y.shape[2]),
order='F')
P = arpfit(Yr, p=2, pixels=active_pixels)
Y_res = Yr - np.dot(Ain, Cin)
model = ProjectedGradientNMF(n_components=1,
init='random',
random_state=0)
model.fit(np.maximum(Y_res, 0))
fin = model.components_.squeeze()
self.Yr, self.Cin, self.fin, self.Ain, self.P, self.Cn = Yr, Cin, fin, Ain, P, Cn
def init_rois(self, n_components=100, show=False):
Ain,Cin,center = greedyROI2d(self.Y, nr=n_components, gSig=[2,2], gSiz=[7,7], use_median=False)
Cn = np.mean(self.Y, axis=-1)
if show:
pl1 = pl.imshow(Cn,interpolation='none')
pl.colorbar()
pl.scatter(x=center[:,1], y=center[:,0], c='m', s=40)
pl.axis((-0.5,self.Y.shape[1]-0.5,-0.5,self.Y.shape[0]-0.5))
pl.gca().invert_yaxis()
active_pixels = np.squeeze(np.nonzero(np.sum(Ain,axis=1)))
Yr = np.reshape(self.Y,(self.Y.shape[0]*self.Y.shape[1],self.Y.shape[2]),order='F')
P = arpfit(Yr, p=2, pixels=active_pixels)
Y_res = Yr - np.dot(Ain,Cin)
model = ProjectedGradientNMF(n_components=1, init='random', random_state=0)
model.fit(np.maximum(Y_res,0))
fin = model.components_.squeeze()
self.Yr,self.Cin,self.fin,self.Ain,self.P,self.Cn = Yr,Cin,fin,Ain,P,Cn
#%% d1,d2,T = np.shape(Y) #%% greedy initialization nr = 30 t1 = time() Ain,Cin,center = greedyROI2d(Y, nr = nr, gSig = [4,4], gSiz = [9,9]) t_elGREEDY = time()-t1 #%% arpfit active_pixels = np.squeeze(np.nonzero(np.sum(Ain,axis=1))) Yr = np.reshape(Y,(d1*d2,T),order='F') p = 2; P = arpfit(Yr,p=2,pixels = active_pixels) #%% nmf Y_res = Yr - np.dot(Ain,Cin) model = ProjectedGradientNMF(n_components=1, init='random', random_state=0) model.fit(np.maximum(Y_res,0)) fin = model.components_.squeeze() #%% update spatial components t1 = time() A,b = update_spatial_components(Yr, Cin, fin, Ain, d1=d1, d2=d2, sn = P['sn']) t_elSPATIAL = time() - t1
#%% plot centers Cn = local_correlations(Y) plt1 = plt.imshow(Cn,interpolation='none') plt.colorbar() plt.scatter(x=center[:,1], y=center[:,0], c='m', s=40) plt.axis((-0.5,d2-0.5,-0.5,d1-0.5)) plt.gca().invert_yaxis() #%% crd = plot_contours(coo_matrix(Ain[:,::-1]),Cn,thr=0.9) #%% active_pixels = np.squeeze(np.nonzero(np.sum(Ain,axis=1))) Yr = np.reshape(Y,(d1*d2,T),order='F') p = 2; P = arpfit(Yr,p=1,pixels = active_pixels) Y_res = Yr - np.dot(Ain,Cin) model = ProjectedGradientNMF(n_components=1, init='random', random_state=0) model.fit(np.maximum(Y_res,0)) fin = model.components_.squeeze() #%% t1 = time() A,b,Cin = update_spatial_components(Yr, Cin, fin, Ain, d1=d1, d2=d2, sn = P['sn'],dist=2,max_size=8,min_size=3) t_elSPATIAL = time() - t1 #%% crd = plot_contours(A,Cn2,thr=0.9,cmap=pl.cm.gray) #%% t1 = time() C,f,Y_res,Pnew = update_temporal_components(Yr,A,b,Cin,fin,ITER=2,deconv_method = 'spgl1') t_elTEMPORAL2 = time() - t1