Convert bounding box annotated files to VOC-like dataset

.travis.yml

@@ -80,6 +80,7 @@ script:
   - labelme --help
   - labelme --version
   - (cd examples/tutorial && rm -rf apc2016_obj3_json && labelme_json_to_dataset apc2016_obj3.json && python load_label_png.py && git checkout -- .)
+  - (cd examples/bbox_detection && rm -rf data_dataset_voc && ./labelme2voc.py labels.txt data_annotated data_dataset_voc && git checkout -- .)
   - (cd examples/semantic_segmentation && rm -rf data_dataset_voc && ./labelme2voc.py labels.txt data_annotated data_dataset_voc && git checkout -- .)
   - (cd examples/instance_segmentation && rm -rf data_dataset_voc && ./labelme2voc.py labels.txt data_annotated data_dataset_voc && git checkout -- .)
   - (cd examples/video_annotation && rm -rf data_dataset_voc && ./labelme2voc.py labels.txt data_annotated data_dataset_voc && git checkout -- .)

examples/bbox_detection/README.md

@@ -8,3 +8,18 @@ labelme data_annotated --labels labels.txt --nodata --autosave
 ```
 
 ![](.readme/annotation.jpg)
+
+
+## Convert to VOC-like Dataset
+
+```bash
+# It generates:
+#   - data_dataset_voc/JPEGImages
+#   - data_dataset_voc/Annotations
+#   - data_dataset_voc/AnnotationsVisualization
+./labelme2voc.py labels.txt data_annotated data_dataset_voc
+```
+
+<img src="data_dataset_voc/JPEGImages/2011_000003.jpg" width="33%" /> <img src="data_dataset_voc/AnnotationsVisualization/2011_000003.jpg" width="33%" />
+
+<i>Fig1. JPEG image (left), Bounding box annotation visualization (right).</i>

examples/bbox_detection/data_dataset_voc/Annotations/2011_000003.xml

@@ -0,0 +1,37 @@
+<annotation>
+  <folder/>
+  <filename>2011_000003.jpg</filename>
+  <database/>
+  <annotation/>
+  <image/>
+  <size>
+    <height>338</height>
+    <width>500</width>
+    <depth>3</depth>
+  </size>
+  <segmented/>
+  <object>
+    <name>person</name>
+    <pose/>
+    <truncated/>
+    <difficult/>
+    <bndbox>
+      <xmin>191</xmin>
+      <ymin>107</ymin>
+      <xmax>313</xmax>
+      <ymax>329</ymax>
+    </bndbox>
+  </object>
+  <object>
+    <name>person</name>
+    <pose/>
+    <truncated/>
+    <difficult/>
+    <bndbox>
+      <xmin>365</xmin>
+      <ymin>83</ymin>
+      <xmax>500</xmax>
+      <ymax>333</ymax>
+    </bndbox>
+  </object>
+</annotation>

examples/bbox_detection/data_dataset_voc/Annotations/2011_000006.xml

@@ -0,0 +1,61 @@
+<annotation>
+  <folder/>
+  <filename>2011_000006.jpg</filename>
+  <database/>
+  <annotation/>
+  <image/>
+  <size>
+    <height>375</height>
+    <width>500</width>
+    <depth>3</depth>
+  </size>
+  <segmented/>
+  <object>
+    <name>person</name>
+    <pose/>
+    <truncated/>
+    <difficult/>
+    <bndbox>
+      <xmin>91</xmin>
+      <ymin>107</ymin>
+      <xmax>240</xmax>
+      <ymax>330</ymax>
+    </bndbox>
+  </object>
+  <object>
+    <name>person</name>
+    <pose/>
+    <truncated/>
+    <difficult/>
+    <bndbox>
+      <xmin>178</xmin>
+      <ymin>110</ymin>
+      <xmax>298</xmax>
+      <ymax>282</ymax>
+    </bndbox>
+  </object>
+  <object>
+    <name>person</name>
+    <pose/>
+    <truncated/>
+    <difficult/>
+    <bndbox>
+      <xmin>254</xmin>
+      <ymin>115</ymin>
+      <xmax>369</xmax>
+      <ymax>292</ymax>
+    </bndbox>
+  </object>
+  <object>
+    <name>person</name>
+    <pose/>
+    <truncated/>
+    <difficult/>
+    <bndbox>
+      <xmin>395</xmin>
+      <ymin>81</ymin>
+      <xmax>447</xmax>
+      <ymax>117</ymax>
+    </bndbox>
+  </object>
+</annotation>

examples/bbox_detection/data_dataset_voc/Annotations/2011_000025.xml

@@ -0,0 +1,49 @@
+<annotation>
+  <folder/>
+  <filename>2011_000025.jpg</filename>
+  <database/>
+  <annotation/>
+  <image/>
+  <size>
+    <height>375</height>
+    <width>500</width>
+    <depth>3</depth>
+  </size>
+  <segmented/>
+  <object>
+    <name>bus</name>
+    <pose/>
+    <truncated/>
+    <difficult/>
+    <bndbox>
+      <xmin>84</xmin>
+      <ymin>20</ymin>
+      <xmax>435</xmax>
+      <ymax>373</ymax>
+    </bndbox>
+  </object>
+  <object>
+    <name>bus</name>
+    <pose/>
+    <truncated/>
+    <difficult/>
+    <bndbox>
+      <xmin>1</xmin>
+      <ymin>99</ymin>
+      <xmax>107</xmax>
+      <ymax>282</ymax>
+    </bndbox>
+  </object>
+  <object>
+    <name>car</name>
+    <pose/>
+    <truncated/>
+    <difficult/>
+    <bndbox>
+      <xmin>409</xmin>
+      <ymin>167</ymin>
+      <xmax>500</xmax>
+      <ymax>266</ymax>
+    </bndbox>
+  </object>
+</annotation>

examples/bbox_detection/data_dataset_voc/class_names.txt

@@ -0,0 +1,21 @@
+_background_
+aeroplane
+bicycle
+bird
+boat
+bottle
+bus
+car
+cat
+chair
+cow
+diningtable
+dog
+horse
+motorbike
+person
+potted plant
+sheep
+sofa
+train
+tv/monitor

examples/bbox_detection/labelme2voc.py

@@ -0,0 +1,127 @@
+#!/usr/bin/env python
+
+from __future__ import print_function
+
+import argparse
+import glob
+import json
+import os
+import os.path as osp
+
+import lxml.builder
+import lxml.etree
+import numpy as np
+import PIL.Image
+
+import labelme
+
+
+def main():
+    parser = argparse.ArgumentParser(
+        formatter_class=argparse.ArgumentDefaultsHelpFormatter)
+    parser.add_argument('labels_file')
+    parser.add_argument('in_dir', help='input dir with annotated files')
+    parser.add_argument('out_dir', help='output dataset directory')
+    args = parser.parse_args()
+
+    if osp.exists(args.out_dir):
+        print('Output directory already exists:', args.out_dir)
+        quit(1)
+    os.makedirs(args.out_dir)
+    os.makedirs(osp.join(args.out_dir, 'JPEGImages'))
+    os.makedirs(osp.join(args.out_dir, 'Annotations'))
+    os.makedirs(osp.join(args.out_dir, 'AnnotationsVisualization'))
+    print('Creating dataset:', args.out_dir)
+
+    class_names = []
+    class_name_to_id = {}
+    for i, line in enumerate(open(args.labels_file).readlines()):
+        class_id = i - 1  # starts with -1
+        class_name = line.strip()
+        class_name_to_id[class_name] = class_id
+        if class_id == -1:
+            assert class_name == '__ignore__'
+            continue
+        elif class_id == 0:
+            assert class_name == '_background_'
+        class_names.append(class_name)
+    class_names = tuple(class_names)
+    print('class_names:', class_names)
+    out_class_names_file = osp.join(args.out_dir, 'class_names.txt')
+    with open(out_class_names_file, 'w') as f:
+        f.writelines('\n'.join(class_names))
+    print('Saved class_names:', out_class_names_file)
+
+    for label_file in glob.glob(osp.join(args.in_dir, '*.json')):
+        print('Generating dataset from:', label_file)
+        with open(label_file) as f:
+            data = json.load(f)
+        base = osp.splitext(osp.basename(label_file))[0]
+        out_img_file = osp.join(
+            args.out_dir, 'JPEGImages', base + '.jpg')
+        out_xml_file = osp.join(
+            args.out_dir, 'Annotations', base + '.xml')
+        out_viz_file = osp.join(
+            args.out_dir, 'AnnotationsVisualization', base + '.jpg')
+
+        img_file = osp.join(osp.dirname(label_file), data['imagePath'])
+        img = np.asarray(PIL.Image.open(img_file))
+        PIL.Image.fromarray(img).save(out_img_file)
+
+        maker = lxml.builder.ElementMaker()
+        xml = maker.annotation(
+            maker.folder(),
+            maker.filename(base + '.jpg'),
+            maker.database(),    # e.g., The VOC2007 Database
+            maker.annotation(),  # e.g., Pascal VOC2007
+            maker.image(),       # e.g., flickr
+            maker.size(
+                maker.height(str(img.shape[0])),
+                maker.width(str(img.shape[1])),
+                maker.depth(str(img.shape[2])),
+            ),
+            maker.segmented(),
+        )
+
+        bboxes = []
+        labels = []
+        for shape in data['shapes']:
+            if shape['shape_type'] != 'rectangle':
+                print('Skipping shape: label={label}, shape_type={shape_type}'
+                      .format(**shape))
+                continue
+
+            class_name = shape['label']
+            class_id = class_names.index(class_name)
+
+            (xmin, ymin), (xmax, ymax) = shape['points']
+            bboxes.append([xmin, ymin, xmax, ymax])
+            labels.append(class_id)
+
+            xml.append(
+                maker.object(
+                    maker.name(shape['label']),
+                    maker.pose(),
+                    maker.truncated(),
+                    maker.difficult(),
+                    maker.bndbox(
+                        maker.xmin(str(xmin)),
+                        maker.ymin(str(ymin)),
+                        maker.xmax(str(xmax)),
+                        maker.ymax(str(ymax)),
+                    ),
+                )
+            )
+
+        captions = [class_names[l] for l in labels]
+        viz = labelme.utils.draw_instances(
+            img, bboxes, labels, captions=captions
+        )
+        PIL.Image.fromarray(viz).save(out_viz_file)
+
+        with open(out_xml_file, 'wb') as f:
+            f.write(lxml.etree.tostring(xml, pretty_print=True))
+
+
+if __name__ == '__main__':
+    main()

labelme/utils/__init__.py

@@ -10,6 +10,7 @@ from .shape import polygons_to_mask
 from .shape import shape_to_mask
 from .shape import shapes_to_label
 
+from .draw import draw_instances
 from .draw import draw_label
 from .draw import label_colormap
 from .draw import label2rgb

labelme/utils/draw.py

@@ -1,8 +1,10 @@
 import io
+import os.path as osp
 
 import numpy as np
 import PIL.Image
 import PIL.ImageDraw
+import PIL.ImageFont
 
 
 def label_colormap(N=256):
@@ -116,3 +118,40 @@ def draw_label(label, img=None, label_names=None, colormap=None, **kwargs):
     out = PIL.Image.open(f).resize(out_size, PIL.Image.BILINEAR).convert('RGB')
     out = np.asarray(out)
     return out
+
+
+def draw_instances(
+    image=None,
+    bboxes=None,
+    labels=None,
+    masks=None,
+    captions=None,
+):
+    import matplotlib
+
+    # TODO(wkentaro)
+    assert image is not None
+    assert bboxes is not None
+    assert labels is not None
+    assert masks is None
+    assert captions is not None
+
+    viz = PIL.Image.fromarray(image)
+    draw = PIL.ImageDraw.ImageDraw(viz)
+
+    font_path = osp.join(
+        osp.dirname(matplotlib.__file__),
+        'mpl-data/fonts/ttf/DejaVuSans.ttf'
+    )
+    font = PIL.ImageFont.truetype(font_path)
+
+    colormap = label_colormap(255)
+    for bbox, label, caption in zip(bboxes, labels, captions):
+        color = colormap[label]
+        color = tuple((color * 255).astype(np.uint8).tolist())
+
+        xmin, ymin, xmax, ymax = bbox
+        draw.rectangle((xmin, ymin, xmax, ymax), outline=color)
+        draw.text((xmin, ymin), caption, font=font)
+
+    return np.asarray(viz)

setup.py

@@ -19,6 +19,7 @@ del here
 
 
 install_requires = [
+    'lxml',
     'matplotlib',
     'numpy',
     'Pillow>=2.8.0',