search

google / ffn

Flood-Filling Networks for instance segmentation in 3d volumes.
Apache License 2.0
316 stars 100 forks source link

Issue with build_coordinates.py #12

Open MatthewBM opened 5 years ago

MatthewBM commented 5 years ago

Hello, I installed ffn and downloaded the sample data without issue. I am trying to train the sample model and got an error.

I ran this:

(ffn) user:~/ffn-master$ runffn.sh

!/bin/bash

python compute_partitions.py \ --input_volume ~/third_party/neuroproof_examples/validation_sample/groundtruth.h5:stack \ --output_volume ~/third_party/neuroproof_examples/validation_sample/af.h5:af \ --thresholds 0.025,0.05,0.075,0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9 \ --lom_radius 24,24,24 \ --min_size 10000 python build_coordinates.py \ --partition_volumes ~/third_party/neuroproof_examples/validation_sample/af.h5:af \ --coordinate_output ~/third_party/neuroproof_examples/validation_sample/tf_record_file \ --margin 24,24,24

compute_partitions.py runs fine, but with build_coordinates I get this:

Traceback (most recent call last): File "build_coordinates.py", line 110, in app.run(main) File "/home/ncmir-lab/.conda/envs/ffn/lib/python3.6/site-packages/absl/app.py", line 300, in run _run_main(main, args) File "/home/ncmir-lab/.conda/envs/ffn/lib/python3.6/site-packages/absl/app.py", line 251, in _run_main sys.exit(main(argv)) File "build_coordinates.py", line 62, in main name, path, dataset = partvol.split(':') ValueError: not enough values to unpack (expected 3, got 2)

MatthewBM commented 5 years ago

I realized the home path doesn't work in this context, I have fixed it by changing: --partition_volumes ~/third_party/neuroproof_examples/validation_sample/af.h5:af to --partition_volumes validation1:/home/user/third_party/neuroproof_examples/validation_sample/af.h5:af

mk123qwe commented 5 years ago

I realized the home path doesn't work in this context, I have fixed it by changing: --partition_volumes ~/third_party/neuroproof_examples/validation_sample/af.h5:af to --partition_volumes validation1:/home/user/third_party/neuroproof_examples/validation_sample/af.h5:af

Can you get the results in the paper?

MatthewBM commented 5 years ago

I didn't reproduce the exact demo data in the paper because it's a different species than what I'm interested in but it performs great on my data.

mk123qwe commented 5 years ago

I didn't reproduce the exact demo data in the paper because it's a different species than what I'm interested in but it performs great on my data.

Run_inference.py can get the result of each 2D segmentation ,but how to reconstruct the result of 3D segmentation?

MatthewBM commented 5 years ago

I'm not sure what you're asking. FFN is a 3D solver, the output is 3D.

mk123qwe commented 5 years ago

I'm not sure what you're asking. FFN is a 3D solver, the output is 3D.

Output shape is (250, 250, 250)，it shows that 250(numbers) x 2D segmentation(w:250,h:250).

mk123qwe commented 5 years ago

I didn't reproduce the exact demo data in the paper because it's a different species than what I'm interested in but it performs great on my data.

HELLO， There is no validation set in the code. How do you determine if your training results are great?

oradomskyi commented 4 years ago

Hi @MatthewBM , @mk123qwe do you know how to run ffn on a dataset of about 5TB?

MatthewBM commented 4 years ago

Hi @FrayaMiner,

Why don't you send me an email matthew.madany@gmail.com with a little more detail about what you're trying to do.