by, bx = (mat1 * mat2).nonzero()
binImage[by, bx, bin] = 1
for binRange in range(-numSigma, numSigma + 1):
curBin = bin + binRange
# Check boundary conditions
if curBin < 0 or curBin >= numBins:
continue
upthresh = stepSize * (curBin + 1)
lowthresh = stepSize * (curBin)
# Calculate cumulative distribution
# Using center of stepsize
xVal = lowthresh + (float(upthresh - lowthresh) / 2)
outmat = (normpdf(xVal, image, stepSize)) * binImage[:, :, bin]
depthMat[:, :, curBin] += outmat[:, :]
# Normalize with max as 1
# depthMat = depthMat / normVal
# Normalize with mean/std
matMean = np.mean(depthMat)
matStd = np.std(depthMat)
depthMat = (depthMat - matMean) * (targetStd / matStd) + targetMean
# Write data for frame
writeData(outMatFile, depthMat, frameIdx)
outMatFile.close()
# y, x, z = depthMat.nonzero()
#
# fig = plt.figure()
# ax = Axes3D(fig)
# ax.scatter(x, z, -y, zdir = 'z', color = 'r')
# plt.show()
curBin = bin + binRange
if curBin < 0 or curBin >= numBins:
continue
upthresh = stepSize * (curBin + 1)
lowthresh = stepSize * (curBin)
#Calculate cumulative distribution
#Using center of stepsize
xVal = lowthresh + (float(upthresh - lowthresh)/2)
outmat = (normpdf(xVal, idxMat, stepSize)) * binMat[:,:,bin]
posMat[:, :, curBin] += outmat[:,:]
#Normalize with mean/std
matMean = np.mean(posMat)
matStd = np.std(posMat)
posMat = (posMat - matMean) * (targetStd/matStd) + targetMean
writeHeaderFile(outMatFile, (Y, X, numBins), 1)
writeData(outMatFile, posMat, 1)
outMatFile.close()
#y, x, z = posMat.nonzero()
#
#fig = plt.figure()
#ax = Axes3D(fig)
#ax.scatter(x, z, -y, zdir = 'z', color = 'r')
#plt.show()
curBin = bin + binRange
if curBin < 0 or curBin >= numBins:
continue
upthresh = stepSize * (curBin + 1)
lowthresh = stepSize * (curBin)
#Calculate cumulative distribution
#Using center of stepsize
xVal = lowthresh + (float(upthresh - lowthresh) / 2)
outmat = (normpdf(xVal, idxMat, stepSize)) * binMat[:, :, bin]
posMat[:, :, curBin] += outmat[:, :]
#Normalize with mean/std
matMean = np.mean(posMat)
matStd = np.std(posMat)
posMat = (posMat - matMean) * (targetStd / matStd) + targetMean
writeHeaderFile(outMatFile, (Y, X, numBins), 1)
writeData(outMatFile, posMat, 1)
outMatFile.close()
#y, x, z = posMat.nonzero()
#
#fig = plt.figure()
#ax = Axes3D(fig)
#ax.scatter(x, z, -y, zdir = 'z', color = 'r')
#plt.show()
#Write out header
#only 1 frame
# writeHeaderFile(outMatFile, (Y, X, numBins), 1)
#for bin in range(numBins):
# binupthresh = stepSize * (bin + 1)
by, bx = (mat1 * mat2).nonzero()
binImage[by, bx, bin] = 1
for binRange in range(-numSigma, numSigma + 1):
curBin = bin + binRange
#Check boundary conditions
if curBin < 0 or curBin >= numBins:
continue
upthresh = stepSize * (curBin + 1)
lowthresh = stepSize * (curBin)
#Calculate cumulative distribution
#Using center of stepsize
xVal = lowthresh + (float(upthresh - lowthresh) / 2)
outmat = (normpdf(xVal, image, stepSize)) * binImage[:, :, bin]
depthMat[:, :, curBin] += outmat[:, :]
#Normalize with max as 1
#depthMat = depthMat / normVal
#Normalize with mean/std
matMean = np.mean(depthMat)
matStd = np.std(depthMat)
depthMat = (depthMat - matMean) * (targetStd / matStd) + targetMean
#Write data for frame
writeData(outMatFile, depthMat, frameIdx)
outMatFile.close()
#y, x, z = depthMat.nonzero()
#
#fig = plt.figure()
#ax = Axes3D(fig)
#ax.scatter(x, z, -y, zdir = 'z', color = 'r')
#plt.show()