davrot/gevi
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Update README.md

Browse source
This commit is contained in:
David Rotermund 2023-07-11 01:37:33 +02:00 committed by GitHub
parent 30a586eeac
commit 5c41e416b8
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 18 additions and 13 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

31
README.md
View file

@ -31,7 +31,7 @@ We used a RTX 3090 as test GPU.
- try to load previous mask - try to load previous mask
- start: cleaned_load_data - start: cleaned_load_data
- start: load_data - start: load_data
- work in XXXX.npy - work on XXXX.npy
- np.load - np.load
- organize acceptor (move to GPU memory) - organize acceptor (move to GPU memory)
- organize donor (move to GPU memory) - organize donor (move to GPU memory)
@ -44,22 +44,23 @@ We used a RTX 3090 as test GPU.
- move inter timeseries - move inter timeseries
- acceptor time series and donor reference image - acceptor time series and donor reference image
- spatial pooling (i.e. 2d average pooling layer) - spatial pooling (i.e. 2d average pooling layer)
- data(x,y,t) = data(x,y,t) / data(x,y,t).mean(t) + 1 - acceptor(x,y,t) = acceptor(x,y,t) / acceptor(x,y,t).mean(t) + 1
- donor(x,y,t) = donor(x,y,t) / donor(x,y,t).mean(t) + 1
- remove the heart beat via SVD from donor and acceptor - remove the heart beat via SVD from donor and acceptor
- copy donor and acceptor and work on the copy with the SVD - copy donor and acceptor and work on the copy with the SVD
- remove the mean (over time) - remove the mean (over time)
- use Cholesky whitening on data with SVD - use Cholesky whitening on data with SVD
- scale the time series accoring the spatial whitening - scale the time series accoring the spatial whitening
- average time series over the spatial dimension (which is the global heart beat) - average time series over the spatial dimension (which is the global heart beat)
- use a normalized scalar product for get spatial scaling factors - use a normalized scalar product for getting spatial scaling factors
- scale the heartbeat with the spatial scaling factors into donor_residuum and acceptor_residuum - scale the heartbeat with the spatial scaling factors into donor_residuum and acceptor_residuum
- store the heartbeat as well as substract it from the original donor and acceptor timeseries - store the heartbeat as well as substract it from the original donor and acceptor timeseries
- remove mean from donor and acceptor timeseries - remove mean from donor and acceptor timeseries (- mean over time)
- remove linear trends from donor and acceptor timeseries - remove linear trends from donor and acceptor timeseries (create a linear function and use a normalized scalar product for getting spatial scaling factors)
- use the SVD heart beat for determining the scaling factors for donor and acceptor (heartbeat_scale) - use the SVD heart beat for determining the scaling factors for donor and acceptor (heartbeat_scale)
- apply bandpass donor_residuum (filtfilt) - apply bandpass donor_residuum (filtfilt)
- apply bandpass acceptor_residuum (filtfilt) - apply bandpass acceptor_residuum (filtfilt)
- a normalized scalar product is used to determine the scale factor scale(x,y) from donor_residuum(x,y,t) and acceptor_residuum(x,y,t) - a normalized scalar product is used to determine the scale factor scale(x,y) between donor_residuum(x,y,t) and acceptor_residuum(x,y,t)
- calculate mask (optional) ; based on the heart beat power at the spatial positions - calculate mask (optional) ; based on the heart beat power at the spatial positions
- scale acceptor signal (heartbeat_scale_a(x,y) * result_a(x,y,t)) and donor signal (heartbeat_scale_d(x,y) * result_d(x,y,t)) - scale acceptor signal (heartbeat_scale_a(x,y) * result_a(x,y,t)) and donor signal (heartbeat_scale_d(x,y) * result_d(x,y,t))
- heartbeat_scale_a = torch.sqrt(scale) - heartbeat_scale_a = torch.sqrt(scale)
@ -74,8 +75,8 @@ We used a RTX 3090 as test GPU.
- try to load previous mask - try to load previous mask
- start cleaned_load_data - start cleaned_load_data
- start load_data - start load_data
- work in XXXX.npy - work on XXXX.npy
- np.load - np.load (load one trial)
- organize acceptor (move to GPU memory) - organize acceptor (move to GPU memory)
- organize donor (move to GPU memory) - organize donor (move to GPU memory)
- organize oxygenation (move to GPU memory) - organize oxygenation (move to GPU memory)
@ -87,17 +88,21 @@ We used a RTX 3090 as test GPU.
- rotate inter timeseries - rotate inter timeseries
- acceptor time series and donor reference image; transformation also used on volume - acceptor time series and donor reference image; transformation also used on volume
- move inter timeseries - move inter timeseries
- acceptor time series and donor reference image; transformation also used on volume - acceptor time series and donor reference image; transformation also used on volume
- spatial pooling (i.e. 2d average pooling layer) - spatial pooling (i.e. 2d average pooling layer)
- data(x,y,t) = data(x,y,t) / data(x,y,t).mean(t) + 1 - acceptor(x,y,t) = acceptor(x,y,t) / acceptor(x,y,t).mean(t) + 1
- donor(x,y,t) = donor(x,y,t) / donor(x,y,t).mean(t) + 1
- oxygenation(x,y,t) = oxygenation(x,y,t) / oxygenation(x,y,t).mean(t) + 1
- volume(x,y,t) = volume(x,y,t) / volume(x,y,t).mean(t) + 1
- frame shift - frame shift
- the first frame of donor and acceptor time series is dropped - the first frame of donor and acceptor time series is dropped
- the oxygenation and volume time series are interpolated between two frames (to compensate for the 5ms delay) - the oxygenation and volume time series are interpolated between two frames (to compensate for the 5ms delay)
- measure heart rate (measure_heartbeat_frequency) i.e. find the frequency with the highest power in the frequency band - measure heart rate (measure_heartbeat_frequency) i.e. find the frequency f_HB(x,y) with the highest power in the frequency band in the volume signal
- use "regression" (i.e. iterative non-orthogonal basis decomposition); remove offset, linear trend, oxygenation and volume timeseries - use "regression" (i.e. iterative non-orthogonal basis decomposition); remove offset, linear trend, oxygenation and volume timeseries
- donor: measure heart beat spectral power (measure_heartbeat_power) - donor: measure heart beat spectral power (measure_heartbeat_power) f_HB(x,y) +/- 3Hz; results in power_d(x,y)
- acceptor: measure heart beat spectral power (measure_heartbeat_power) - acceptor: measure heart beat spectral power (measure_heartbeat_power) f_HB(x,y) +/- 3Hz ; results in power_a(x,y)
- scale acceptor and donor signals via the powers - scale acceptor and donor signals via the powers
- scale(x,y) = power_d(x,y) / (power_a(x,y) + 1e-20)
- heartbeat_scale_a = torch.sqrt(scale) - heartbeat_scale_a = torch.sqrt(scale)
- heartbeat_scale_d = 1.0 / (heartbeat_scale_a + 1e-20) - heartbeat_scale_d = 1.0 / (heartbeat_scale_a + 1e-20)
- result(x,y,t) = 1.0 + result_a(x,y,t) - result_d(x,y,t) - result(x,y,t) = 1.0 + result_a(x,y,t) - result_d(x,y,t)