1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
|
# Semantic Segmentation Example
## Annotation
```bash
labelme data_annotated --labels labels.txt --nodata --validatelabel exact --config '{shift_auto_shape_color: -2}'
```
## Convert to VOC-format Dataset
```bash
# It generates:
# - data_dataset_voc/JPEGImages
# - data_dataset_voc/SegmentationClass
# - data_dataset_voc/SegmentationClassNpy
# - data_dataset_voc/SegmentationClassVisualization
./labelme2voc.py data_annotated data_dataset_voc --labels labels.txt --noobject
```
<img src="data_dataset_voc/JPEGImages/2011_000003.jpg" width="33%" /> <img src="data_dataset_voc/SegmentationClass/2011_000003.png" width="33%" /> <img src="data_dataset_voc/SegmentationClassVisualization/2011_000003.jpg" width="33%" />
Fig 1. JPEG image (left), PNG label (center), JPEG label visualization (right)
Note that the label file contains only very low label values (ex. `0, 4, 14`), and
`255` indicates the `__ignore__` label value (`-1` in the npy file).
You can see the label PNG file by following.
```bash
labelme_draw_label_png data_dataset_voc/SegmentationClass/2011_000003.png
```
<img src=".readme/draw_label_png.jpg" width="33%" />
|
