Angelou0516 commited on
Commit
2a3aff4
·
verified ·
1 Parent(s): 95df323

Upload CREMI dataset (ssTEM neuron segmentation)

Browse files
Files changed (5) hide show
  1. .gitattributes +3 -0
  2. README.md +79 -0
  3. train/sample_A.hdf +3 -0
  4. train/sample_B.hdf +3 -0
  5. train/sample_C.hdf +3 -0
.gitattributes CHANGED
@@ -57,3 +57,6 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
57
  # Video files - compressed
58
  *.mp4 filter=lfs diff=lfs merge=lfs -text
59
  *.webm filter=lfs diff=lfs merge=lfs -text
 
 
 
 
57
  # Video files - compressed
58
  *.mp4 filter=lfs diff=lfs merge=lfs -text
59
  *.webm filter=lfs diff=lfs merge=lfs -text
60
+ train/sample_A.hdf filter=lfs diff=lfs merge=lfs -text
61
+ train/sample_B.hdf filter=lfs diff=lfs merge=lfs -text
62
+ train/sample_C.hdf filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,79 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # CREMI - Circuit Reconstruction from Electron Microscopy Images
2
+
3
+ ## Dataset Description
4
+
5
+ CREMI is a benchmark dataset for evaluating algorithms for automatic reconstruction of neurons and neuronal connectivity from serial section electron microscopy data.
6
+
7
+ ### Modality
8
+ - Serial section Transmission Electron Microscopy (ssTEM)
9
+
10
+ ### Anatomy
11
+ - Adult Drosophila melanogaster brain
12
+
13
+ ### Segmentation Targets
14
+ - **Neuron segmentation** (`volumes/labels/neuron_ids`)
15
+ - **Synaptic cleft segmentation** (`volumes/labels/clefts`)
16
+
17
+ ### Volume Size
18
+ - 125 slices × 1250 × 1250 pixels per sample
19
+ - Resolution: 4×4×40 nm (x,y,z)
20
+
21
+ ## Dataset Structure
22
+
23
+ ```
24
+ CREMI/
25
+ └── train/
26
+ ├── sample_A.hdf # 167 MB
27
+ ├── sample_B.hdf # 160 MB
28
+ └── sample_C.hdf # 165 MB
29
+ ```
30
+
31
+ Note: Test files (A+, B+, C+) are not included as they have no labels.
32
+
33
+ ## HDF5 Structure
34
+
35
+ Each training file contains:
36
+ - `volumes/raw`: (125, 1250, 1250) uint8 - grayscale EM images
37
+ - `volumes/labels/neuron_ids`: (125, 1250, 1250) uint64 - neuron instance segmentation
38
+ - `volumes/labels/clefts`: (125, 1250, 1250) uint64 - synaptic cleft segmentation
39
+ - `annotations/` - synaptic partner annotations
40
+
41
+ ## Usage
42
+
43
+ ```python
44
+ import h5py
45
+ from huggingface_hub import hf_hub_download
46
+
47
+ # Download a training volume
48
+ path = hf_hub_download(
49
+ repo_id="Angelou0516/CREMI",
50
+ filename="train/sample_A.hdf",
51
+ repo_type="dataset"
52
+ )
53
+
54
+ # Load data
55
+ with h5py.File(path, 'r') as f:
56
+ raw = f['volumes/raw'][:] # (125, 1250, 1250)
57
+ neuron_ids = f['volumes/labels/neuron_ids'][:] # (125, 1250, 1250)
58
+ clefts = f['volumes/labels/clefts'][:] # (125, 1250, 1250)
59
+ ```
60
+
61
+ ## Citation
62
+
63
+ ```bibtex
64
+ @misc{cremi2016,
65
+ title={CREMI: MICCAI Challenge on Circuit Reconstruction from Electron Microscopy Images},
66
+ author={Funke, Jan and Saalfeld, Stephan and Bock, Davi and Turaga, Srini and Perlman, Eric},
67
+ year={2016},
68
+ url={https://cremi.org/}
69
+ }
70
+ ```
71
+
72
+ ## License
73
+
74
+ Please refer to the original dataset website: https://cremi.org/
75
+
76
+ ## Links
77
+ - [CREMI Challenge Website](https://cremi.org/)
78
+ - [CREMI Data Page](https://cremi.org/data/)
79
+ - [CREMI Metrics](https://cremi.org/metrics/)
train/sample_A.hdf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4c563d1b78acb2bcfb3ea958b6fe1533422f7f4a19f3e05b600bfa11430b510d
3
+ size 175159177
train/sample_B.hdf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:887e85521e00deead18c94a21ad71f278d88a5214c7edeed943130a1f4bb48b8
3
+ size 167293425
train/sample_C.hdf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2874496f224d222ebc29d0e4753e8c458093e1d37bc53acd1b69b19ed1ae7052
3
+ size 173102370