Skip to content

Commit

Permalink
Merge pull request #55 from opendatalab/ling
Browse files Browse the repository at this point in the history 
docs: update docs
  • Loading branch information
@lvlinsheng
lvlinsheng authored Nov 16, 2023
2 parents d71bd08 + 84b1c54 commit 88f1777
Show file tree
Hide file tree
Showing 12 changed files with 234 additions and 214 deletions.
10 changes: 5 additions & 5 deletions src/pages/guide.account/markdown_en-US.mdx
View file
Original file line number Diff line number Diff line change
@@ -1,17 +1,17 @@
import xlabAcount from './xlab-account-cn.png';
import localAcount from './local-account-cn.png';

## Account
## Registration and Login

### 在线版
### Web online version

可通过“手机验证码”注册登录，也可使用“手机号/邮箱+密码”的方式注册登录。
You can register and log in using "Mobile Verification Code", or you can also register and log in with "Mobile Number/Email + Password".

<img src={xlabAcount} />

### 离线版
### Local deployment version

账号信息存储在本地电脑，可通过用邮箱注册来进行登录。
You can log in by registering with an email.

<img src={localAcount} />

35 changes: 28 additions & 7 deletions src/pages/guide.export/markdown_en-US.mdx
View file
Original file line number Diff line number Diff line change
Expand Up @@ -2,22 +2,43 @@ import imageUrl from './image.png';
import video from './video.png';
import audio from './audio.png';

## 导出标注结果
## Export Annotated Results

### Image Format

After finishing the annotations, you can export the annotated results file in JSON, COCO, MASK formats.For annotation format instructions, please visit https://opendatalab.github.io/labelU/#/schema/image/point.






### 图片

完成标注后，可将标注结果文件以JSON,COCO,MASK形式导出。

<img src={imageUrl} />

### 视频
### Video Format

After completing the annotations, you can export the annotated results file in JSON format.For annotation format instructions, please visit https://opendatalab.github.io/labelU/#/schema/video/segment.







完成标注后，可将标注结果文件以JSON形式导出。

<img src={video} />

### 音频
### Audio Format

After completing the annotations, you can export the annotated results file in JSON format.For annotation format instructions, please visit https://opendatalab.github.io/labelU/#/schema/audio/segment.







完成标注后，可将标注结果文件以JSON形式导出。

<img src={audio} />
6 changes: 3 additions & 3 deletions src/pages/guide.export/markdown_zh-CN.mdx
View file
Original file line number Diff line number Diff line change
Expand Up @@ -6,18 +6,18 @@ import audio from './audio.png';

### 图片

完成标注后，可将标注结果文件以JSON,COCO,MASK形式导出。
完成标注后，可将标注结果文件以JSON,COCO,MASK形式导出。标注格式说明见https://opendatalab.github.io/labelU/#/schema/image/point

<img src={imageUrl} />

### 视频

完成标注后，可将标注结果文件以JSON形式导出。
完成标注后，可将标注结果文件以JSON形式导出。标注格式说明见https://opendatalab.github.io/labelU/#/schema/video/segment

<img src={video} />

### 音频

完成标注后，可将标注结果文件以JSON形式导出。
完成标注后，可将标注结果文件以JSON形式导出。标注格式说明见https://opendatalab.github.io/labelU/#/schema/audio/segment

<img src={audio} />
48 changes: 24 additions & 24 deletions src/pages/guide.introduction/markdown_en-US.mdx
View file
Original file line number Diff line number Diff line change
@@ -1,49 +1,49 @@
## 产品简介
## Product Introduction

### 产品介绍
### Product Overview

LabelU是一个开源的数据标注工具，它可以帮助用户快速、准确、高效地对数据进行标注，从而提高机器学习模型的性能和质量。LabelU支持多种标注类型，包括标签分类、拉框、多边形、点、线、文本描述等，满足不同场景和需求的标注任务。
体验产品可通过以下两种方式：
- 在线体验：[https://labelu.shlab.tech/](https://labelu.shlab.tech/)
- 本地部署：[https://opendatalab.github.io/labelU/#/guide/install](https://opendatalab.github.io/labelU/#/guide/install)
LabelU is an open-source data annotation tool that can help users quickly, accurately, and efficiently annotate data, thereby improving the performance and quality of machine learning models. LabelU supports various types of annotations, including label classification, bounding boxes, polygons, points, lines, text descriptions, etc., meeting different scenarios and annotation task needs.
You can experience the product in two ways:
- Online experience：[https://labelu.shlab.tech/](https://labelu.shlab.tech/)
- Local deployment：[https://opendatalab.github.io/labelU/#/guide/install](https://opendatalab.github.io/labelU/#/guide/install)

### 功能特性
### Feature Introduction

LabelU提供了多种标注工具和功能，支持图像、视频、音频标注。
LabelU provides a variety of annotation tools and features, supporting image, video, and audio annotations.

- 图像类：多功能图像处理工具，涵盖2D框、语义分割、多段线、关键点等多种标注工具，协助完成图像的标识、注释和分析。
- 视频类：具备强大视频处理能力，可实现视频分割、视频分类、视频信息提取等功能，为模型训练提供优质标注数据。
- 音频类：高效精准的音频分析工具，可实现音频分割、音频分类、音频信息提取等功能，将复杂的声音信息直观可视化。
- Image-based: Multifunctional image processing tools, covering 2D box, semantic segmentation, polyline, key points, and various other annotation tools，to assist in the identification, annotation, and analysis of images.
- Video-based: With powerful video processing capabilities, it can perform video segmentation, video classification, video information extraction, and other functions, providing high-quality annotated data for model training.
- Audio-based: Efficient and accurate audio analysis tools, capable of audio segmentation, audio classification, audio information extraction, and other functions, making complex sound information intuitively visualized.

### 概念说明
### Concept Introduction

<table>
<thead>
<tr>
<th>名词</th>
<th>说明</th>
<th>Concept</th>
<th>Explanation</th>
</tr>
</thead>
<tbody>
<tr>
<td style={{ whiteSpace: 'nowrap' }}>任务（task）</td>
<td>为了对某个数据集进行标注而建立的任务</td>
<td style={{ whiteSpace: 'nowrap' }}>Task</td>
<td>A task established for annotating a specific dataset</td>
</tr>
<tr>
<td style={{ whiteSpace: 'nowrap' }}>标签（label）</td>
<td>标注时需要添加的分类标识，比如猫、狗、行人、车辆</td>
<td style={{ whiteSpace: 'nowrap' }}>Label</td>
<td>The classification labels that need to be added during annotation, such as cat, dog, pedestrian, vehicle</td>
</tr>
<tr>
<td style={{ whiteSpace: 'nowrap' }}>标记（annotation）</td>
<td>进行一次标注后生成的对象，比如一个矩形框、一个点</td>
<td style={{ whiteSpace: 'nowrap' }}>Annotation</td>
<td>The object generated after a round of annotation, like a rectangle box, a point</td>
</tr>
<tr>
<td style={{ whiteSpace: 'nowrap' }}>属性（attribute）</td>
<td>对标签的进一步描述，比如将某个物体标为车辆后，添加属性“车辆遮挡率为20%”</td>
<td style={{ whiteSpace: 'nowrap' }}>Attribute</td>
<td>Further description of the label, for example, after labeling an object as a vehicle, adding the attribute "vehicle occlusion rate is 20%"</td>
</tr>
<tr>
<td style={{ whiteSpace: 'nowrap' }}>标注结果（result）</td>
<td>标记+标签+属性，一条完整的标注记录</td>
<td style={{ whiteSpace: 'nowrap' }}>Result</td>
<td>Annotation + Label + Attribute, a complete annotation record</td>
</tr>
</tbody>
</table>
Expand Down
34 changes: 17 additions & 17 deletions src/pages/guide.task-annotation.audio/markdown_zh-CN.mdx
View file
Original file line number Diff line number Diff line change
@@ -1,44 +1,44 @@
import global from './global.png';
import tool from './tool.png';

## 音频标注
## Audio Annotation

标注流程说明：
Annotation process description：

1. 判断任务是否为无效，若无效点击跳过，进入下一音频，若有效，有标注任务时（目标检测、语意分割、线标注、点标注），标注工具和配置的一致
2. 在全局标签上，可以勾选预先配置好的「标签分类」来给整个音频打上标签，也可以在「文本描述」下对于视频进行文字说明
3. 在具体标记上，首先点击左上角的「工具样式」来选择标记方式，在音频中完成该样式的标记后同步在右方的标记栏中。对于该标签的名称等信息可以手动点击后进行修改
4. 在右侧标签结果管理栏单击标签结果可选中音频中对应标记，工具栏中选中工具切换为标签结果的工具。点击「修改」详细信息可下拉选择标签名称，点击「隐藏」按钮可隐藏这一条标记信息，点击「删除」按钮可删除这一条标记信息
5. 选择「下一页」，进入下一个音频
6. 重复1~5，直到标注完成
1. Determine whether the task is invalid; if so, click skip to proceed to the next audio. If valid, when there are annotation tasks (object detection, semantic segmentation, line annotation, point annotation), the annotation tools and configuration are consistent.
2. On the global labels, you can select the pre-configured "label classification" to tag the entire audio or provide a text explanation for the audio under "text description".
3. For specific markings, first click the "tool style" in the upper left corner to select the marking method. After completing the style of the mark in the audio, synchronize it in the label column on the right. The name of this label and other information can be modified manually by clicking on it.
4. Click on the label result in the label result management column on the right side to select the corresponding marking in the audio. The selected tool in the toolbar switches to the label result tool. Click "modify" for detailed information to drop down and select the label name; click the "hide" button to hide this marking information; click the "delete" button to delete this marking information.
5. Choose "Next", go to the next audio.
6. Repeat steps 1-5 until the annotation is complete.

### 工具介绍
### Tool introduction

<table>
<thead>
<tr>
<th style={{ whiteSpace: 'nowrap' }}>工具样式</th>
<th>使用方法</th>
<th style={{ whiteSpace: 'nowrap' }}>Tool Style</th>
<th>Usage Method</th>
</tr>
</thead>
<tbody>
<tr>
<td>片段分割</td>
<td>播放音频并找到你想要开始切割的点，按下暂停并点击确定起始点。可以直接在时间线上点击并拖动选择终止点，来选择你想要切割的音频部分。</td>
<td>Segment Segmentation</td>
<td>Play the audio and find the point where you want to start cutting, pause and click to confirm the starting point. You can directly click and drag to select the end point on the timeline, to choose the part of the audio you want to cut.</td>
</tr>
<tr>
<td>时间戳</td>
<td>选中找到你想引用或高亮的时间点，如果你想标记音频的1小时10分钟30秒处，你应该点击01:10:30这个进度条的点。</td>
<td>Timestamps</td>
<td>Select the time point you want to reference or highlight. If you want to mark 1 hour, 10 minutes and 30 seconds of the audio, you should click the point on the progress bar at 01:10:30.</td>
</tr>
</tbody>
</table>


### 全局标签
### Global labels

<img src={global} />


### 标记
### Specific markings

<img src={tool} />
46 changes: 23 additions & 23 deletions src/pages/guide.task-annotation.image/markdown_en-US.mdx
View file
Original file line number Diff line number Diff line change
@@ -1,55 +1,55 @@
import global from './global.png';
import tool from './tool.png';

## 图片标注
## Image Annotation

标注流程说明：
Annotation process description:

1. 判断任务是否为无效，若无效点击跳过，进入下一张图片，若有效，有标注任务时（目标检测、语意分割、线标注、点标注），标注工具和配置的一致
2. 在全局标签上，可以勾选预先配置好的「标签分类」来给整个图片打上标签，也可以在「文本描述」下对于图片进行文字说明
3. 在具体标记上，首先点击左上角的「工具样式」来选择标记方式，在图中完成该样式的标记后同步在右方的标记栏中。对于该标签的名称等信息可以手动点击后进行修改
4. 在右侧标签结果管理栏单击标签结果可选中图片中对应标记，工具栏中选中工具切换为标签结果的工具。点击「修改」详细信息可下拉选择标签名称，点击「隐藏」按钮可隐藏这一条标记信息，点击「删除」按钮可删除这一条标记信息
5. 选择「下一页」，进入下一个图片
6. 重复1~5，直到标注完成
1. Determine whether the task is invalid; if so, click skip to proceed to the next image. If valid, when there are annotation tasks (object detection, semantic segmentation, line annotation, point annotation), the annotation tools and configuration are consistent.
2. On the global labels, you can select the pre-configured "label classification" to tag the entire image or provide a text explanation for the image under "text description".
3. For specific markings, first click the "tool style" in the upper left corner to select the marking method. After completing the style of the mark in the image, synchronize it in the label column on the right. The name of this label and other information can be modified manually by clicking on it.
4. Click on the label result in the label result management column on the right to select the corresponding mark in the image. The selected tool in the toolbar switches to the label result tool. Click "modify" for detailed information to drop down and select the label name; click the "hide" button to hide this marking information; click the "delete" button to delete this marking information.
5. Choose "Next", go to the next image.
6. Repeat steps 1-5 until the annotation is complete.

### 全局标签
### Global labels

<img src={global} />


### 标记
### Specific markings:

<img src={tool} />

### 工具介绍
### Tool introduction

<table>
<thead>
<tr>
<th>工具样式</th>
<th style={{ whiteSpace: 'nowrap' }}>使用方法</th>
<th>Tool Style</th>
<th style={{ whiteSpace: 'nowrap' }}>Usage Method</th>
</tr>
</thead>
<tbody>
<tr>
<td>拉框标注</td>
<td>1. 选中拉框工具并配置标签，如轿车、公交车。<br />2. 单击鼠标左键标注第一点，画出框的范围之后再次单击左键，即可绘制出框。<br />3. 右键选中框之后，可以调整框的大小、有效性，也可以删除已标注的框。</td>
<td>Draw-box annotation</td>
<td>1. Select the draw-box tool and configure the labels, such as cars, buses.<br />2. Click the left mouse button to annotate the first point, draw out the range of the box and click the left button again to draw the box. <br />3. After right-clicking to select the box, you can adjust the size and validity of the box or delete the annotated box.</td>
</tr>
<tr>
<td>标点标注</td>
<td>1. 选中标点工具并配置标签，如人体姿态14个关键点，包括头、脖子、左肩、右肩、左手肘、右手肘、左腕、右腕、左髋、右髋、左膝、右膝、左脚踝、右脚踝。<br />2. 单击鼠标左键标注指定关键点。<br />3. 右键选中点之后，可以调整点的位置、属性，也可以删除已标注的点。</td>
<td>Point annotation</td>
<td>1. Select the point tool and configure the labels, such as 14 key points of human posture, including head, neck, left shoulder, right shoulder, left elbow, right elbow, left wrist, right wrist, left hip, right hip, left knee, right knee, left ankle, right ankle. <br />2. Click the left mouse button to annotate the specified key points.<br />3. After right-clicking to select a point, you can adjust the position and attributes of the point or delete the annotated point.</td>
</tr>
<tr>
<td style={{ whiteSpace: 'nowrap' }}>多边形标注</td>
<td>1. 选中多边形工具并配置标签，如猫、狗。<br />2. 单击鼠标左键标注起始点，随后沿目标边缘多次单击左键，以此类推在围绕目标边缘右键连接最接近起始点后，即可绘制出多边形框。<br />3. 右键选中框之后，可以调整目标边缘关键点、线段、目标有效性，也可以删除已标注的多边形框。</td>
<td style={{ whiteSpace: 'nowrap' }}>Polygon annotation</td>
<td>1. Select the polygon tool and configure the labels, such as cats, dogs. <br />2. Click the left mouse button to annotate the starting point, then click the left button multiple times along the target edge. After right-clicking close to the starting point around the target edge, you can draw a polygon frame. <br />3. After right-clicking to select the frame, you can adjust the key points of the target edge, line segments, target validity or delete the annotated polygon frame.</td>
</tr>
<tr>
<td>标线标注</td>
<td>1. 选中线条工具并配置标签，如车道线。<br />2. 单击鼠标左键标注起始点，再次点击1+N后右键为终止点。<br />3. 操作Shift+左键则为垂直或水平线。</td>
<td>Line annotation</td>
<td>1. Select the line tool and configure the labels, such as lane lines.<br />2. Click the left mouse button to annotate the starting point, then click again after 1+N for the end point. <br />3. Operating Shift + left button results in a vertical or horizontal line.</td>
</tr>
</tbody>
</table>


**注：** 在标记的过程中，可以点击「撤销」回退至上一步，也可以开启「显示顺序」按钮来看到标记的顺序，针对图片的调整也可以点击左下方的「图片调整」按钮来调整图片的饱和度，对比度，曝光度这些数值。
以上操作都可以通过快捷键实现，具体的操作可以点击左上方「快捷键」按钮来进行查看。
**Note：** In the process of marking, you can click "undo" to return to the previous step, you can also turn on the "display order" button to see the order of marking. For image adjustment, you can click the "image adjustment" button at the bottom left to adjust values such as saturation, contrast, and exposure.
All of the above operations can be implemented through shortcut keys. The specific operation can be viewed by clicking the "shortcut" button in the upper left corner.
Loading

0 comments on commit 88f1777

Please sign in to comment.