Labeling
约 1 分钟阅读
标签:Dataset
- VOC - Visual Object Classes
- Pascal VOC
- XML
- object, name, bndbox
- COCO
- Common Objects in Context
- JSON
- images, annotations, categories
- bbox -
[x, y, w, h]
- YOLO - You Only Look Once
- txt
- VGG Image Annotator (VIA)
- 参考
- datumaro笔记datumaroopenvinotoolkit/datumaro · MIT, Python · 注意⚠️ 依赖的 openvino 可能对 Python 版本有要求,安装不上会导致安装 0.5.0 的 datumaro · openvino==2023.2打开:/notes/ai/ml/datumaro
| format | spec | adopted by |
|---|---|---|
| xyxy | left, top, right, bottom | Pascal VOC |
| xyxyn | Albumentations | |
| xywh | left, top, width, height | COCO |
| xywhn | ||
| cxcywh | center x, center y, width, height | CreateML |
| cxcywhn | YOLO |
- n -> Normalized
YOLO
YOLO Detection
class_index x_center y_center width heightYOLO OOB
class_index, x1, y1, x2, y2, x3, y3, x4, y4- OOB Oriented Bounding Box
YOLO Segmentation
class-index x1 y1 x2 y2 ... xn ynYOLO Pose Estimation 2D
class-index x y width height px1 py1 px2 py2 ... pxn pynYOLO Pose Estimation 3D
class-index x y width height px1 py1 p1-visibility px2 py2 p2-visibility ... pxn pyn pn-visibilityVOC
<annotation>
<folder>VOC2012</folder>
<filename>image1.jpg</filename>
<size>
<width>800</width>
<height>600</height>
<depth>3</depth>
</size>
<object>
<name>dog</name>
<pose>Unspecified</pose>
<truncated>0</truncated>
<difficult>0</difficult>
<bndbox>
<xmin>48</xmin>
<ymin>240</ymin>
<xmax>195</xmax>
<ymax>371</ymax>
</bndbox>
</object>
</annotation>
COCO
{
"images": [
{
"id": 1,
"file_name": "image1.jpg",
"width": 800,
"height": 600
}
],
"annotations": [
{
"id": 1,
"image_id": 1,
"category_id": 18,
"bbox": [48, 240, 147, 131],
"segmentation": [],
"area": 19257,
"iscrowd": 0
}
],
"categories": [
{
"id": 18,
"name": "dog",
"supercategory": "animal"
}
]
}关联信息
反向链接和本文引用的外部资料。
反向链接
- 机器学习笔记 · 标记
训练 · 标记 · Comparing Deep Learning Frameworks · | - | tiny-cnn | caffe | Theano | TensorFlow | · | Prerequisites | Nothing(Optional:TBB,OpenMP) | BLAS,Boost,protobuf,glog,gflags,...
- ML Awesome笔记 · Labeling
LLM · LLM Awesome · Diffusion · Diffusion Awesome · NLP · NLP Awesome · GAN · img2img · Framework · PyTorch · by Meta · tinygrad/tinygrad
References
GitHub
1 条- KKKSQJ/DeepLearning/others/label_convertgithub.com/KKKSQJ/DeepLearning/tree/master/others/label_convert
其他外链
2 条- dragoneye.ai/blog/a-guide-to-bounding-box-formatsdragoneye.ai/blog/a-guide-to-bounding-box-formats
- openvinotoolkit.github.io/datumaro/latest/docs/data-formats/formats/index.htmlopenvinotoolkit.github.io/datumaro/latest/docs/data-formats/formats/index.html