Skip to content

Latest commit

 

History

History
88 lines (59 loc) · 4.35 KB

notes.md

File metadata and controls

88 lines (59 loc) · 4.35 KB

BEADS

precision and recall

1.4mu in xy

01highc_f4 applied to 01highc/4mu: /g/kreshuk/LF_computed/lnet/logs/beads/z_out51/01highc_f4/20-04-29_17-08-09 /g/kreshuk/LF_computed/lnet/logs/beads/01highc_f4/20-04-21_10-49-27/train/checkpoints/v1_checkpoint_SmoothL1Loss=-0.00016112386947497725.pth 4mu_f4 applied to 4mu/01highc: /g/kreshuk/LF_computed/lnet/logs/beads/z_out51/4mu_f4/20-04-29_17-09-19 01highc_f4_lr applied to 01highc/4mu: /g/kreshuk/LF_computed/lnet/logs/beads/z_out51/01highc_f4_lr/20-04-29_17-55-30

0.7mu in xy

01highc_f8 applied to 01highc/4mu: /g/kreshuk/LF_computed/lnet/logs/beads/z_out51/01highc_f8/20-04-29_17-09-06 # THE beads net (beads/01highc/20-04-21_11-41-43/train/checkpoints/v1_checkpoint_SmoothL1Loss=-0.00012947025970788673.pth) /g/kreshuk/LF_computed/lnet/logs/beads/01highc/20-04-21_11-41-43/train/checkpoints/v1_checkpoint_SmoothL1Loss=-0.00012947025970788673.pth

4mu_f8 applied to 4mu/01highc: /g/kreshuk/LF_computed/lnet/logs/beads/z_out51/4mu_f8/20-04-29_17-10-01 f8_lr applied to 01highc: /g/kreshuk/LF_computed/lnet/logs/beads/z_out51/01highc_f8_lr/20-04-29_18-38-22

HEART

best static: /g/kreshuk/LF_computed/lnet/logs/heart2/test_z_out49/f4/z_out49/static_f4_b2_with5_pois/20-06-13_20-26-33/train2/v1_checkpoint_498_MS_SSIM=0.9710696664723483 /g/kreshuk/LF_computed/lnet/logs/heart2/z_out49/static_f4_b2_with5_pois/20-06-13_20-26-33/train2/run000/checkpoints/v1_checkpoint_498_MS_SSIM=0.9710696664723483.pth

best static best refined: /g/kreshuk/LF_computed/lnet/logs/heart2/z_out49/f4_b2_refine_pois/20-06-16_16-35-51/train/run000/checkpoints/v1_checkpoint_155000_MS_SSIM=0.9682732355651673.pth best dynamic training: /g/kreshuk/LF_computed/lnet/logs/heart2/z_out49/f4_b2/20-05-20_10-18-11/train/run000/checkpoints/v1_checkpoint_MSSSIM=0.6722144321961836.pth

static_f4_b2_with5_pois_dir = 'M:\lnet\logs\heart2\test_z_out49\f4\z_out49\static_f4_b2_with5_pois\20-06-13_20-26-33\train2\v1_checkpoint_498_MS_SSIM=0.9710696664723483'; refined_network_dir = 'M:\lnet\logs\heart2\test_z_out49\f4\z_out49\f4_b2_refine_pois\20-06-16_16-35-51\v1_checkpoint_57250_MS_SSIM=0.9605052428382436'; lr_dir = 'M:\lnet\logs\heart2\test_z_out49\lr_f4'; purely_dyn_network_dir = 'M:\lnet\logs\heart2\test_z_out49\f4\z_out49\f4_b2\20-05-20_10-18-11\v1_checkpoint_MSSSIM=0.6722144321961836';

dyn_on_stat: from scratch: /g/kreshuk/beuttenm/pycharm_tmp/repos/hylfm-net/logs/train/heart/fake_dyn/from_scratch/21-01-12_21-05-48/train_01/run000/checkpoints/v1_checkpoint_27500_ms_ssim-scaled=0.8430055000565269.pth

BRAIN

best: K:\LF_computed\lnet\logs\brain1\test_z_out49\f4\z_out49\f4_b2_only11_2\20-06-06_17-59-42\v1_checkpoint_37500_MS_SSIM=0.8822072593371073

Cropping

rect_LF = [174, 324, 1700, 1400]; %[xmin, ymin, width, height]; -> after this crop, rectified rect_LS = [124, 274, 1800, 1500] %[xmin, ymin, width, height];

for data: \TestOutputGcamp\LenseLeNet_Microscope\20200311_Gcamp\fish2\10Hz\slideThrough extra crop on rectified: [323, 133, 1273, 988] %[xmin, ymin, width, height]

20200311 (best day) fish2 (best fish) 5 Hz 10 Hz

10Hz

fish 1 of day 9

fish 4 of day 9

TestOutputGcamp\LenseLeNet_Microscope\20200309_Gcamp\fish4_promising\longRun\fov2\2020-03-09_09.06.55\stack_28_channel_4\SwipeThrough_-450_-210_nimages_49\TP_00000\RC_rectified

not all planes: TestOutputGcamp\LenseLeNet_Microscope\20200309_Gcamp\fish4_promising\fov/single_planes/fov1

all 25: TestOutputGcamp\LenseLeNet_Microscope\20200309_Gcamp\fish4_promising\fov/single_planes/fov2

fish 2

80 x 121 planes (0-121) (possibly filter) TestOutputGcamp\LenseLeNet_Microscope\20200311_Gcamp\fish2\10Hz\2020-03-11_07.34.47\stack_33_channel_3\SwipeThrough_-390_-270_nimages_121

31+15 x 121 planes (60-181) TestOutputGcamp\LenseLeNet_Microscope\20200311_Gcamp\fish2\10Hz\241Planes

7 * 8 x 121 planes TestOutputGcamp\LenseLeNet_Microscope\20200311_Gcamp\fish2\10Hz\slideThrough

fish 4 single planes

resolution

gcamp: day11, 10Hz, z: ls: 121um in 1um steps, 5um, -> z_out: 25 xy: ls: 6.5/22.2=.29 ... -> 2.8um (factor 2)

fish 1 single planes

L:\LenseLeNet_Microscope\20200309_Gcamp\fish1_awesome\2020-03-09_04.35.55

predictions

preliminary volume prediction 1.3mu xy: /g/kreshuk/LF_computed/lnet/logs/brain/z_out51/f4_vol/20-04-28_16-44-06/predict_volume/run000