First Steps

Let's join us!

Create an account

To create an account:

1. Go to register page.

2. Click "Create new account".

3. Fill in the form with your information then click "Sign up".

4. Go to your email and find the confirmation link.

5. Choose "Click to log in".

6. Fill in your new username, password then click "Sign in".

Create a new team

1. Click in your name on the top right of the screen, choose "Teams".

2. In "Team Management", choose "+Add Team".

3. In "Create Team" you can fill in your team information.

4. In "Team members", please add your team here with their roles, you can delete or duplicate team member.

5. Click "Submit".

Create a new dataset

1. Go to "Datasets".

2. Click "Add dataset".

3. Fill in name, description, needed tags, types of data, labels and attributes. Follow guidelines and be specific with the requirements!

4. Upload Dataset by clicking "Click to upload".

5. Click "Submit".

With Labels, you can add attributes and change colors.

Types of the dataset:

• Instance Segmentation: is a computer vision task for detecting and localizing an object in an image. Instance segmentation is a natural sequence of semantic segmentation.

• Object Detection: is a computer vision technique for locating instances of objects in images or videos

• Sematic Segmentation: is the task of assigning a class label to every pixel in the image.

Note:

With 3D data, compress your data with this structure.

Arbitrary folder names are not supported, so folder names must follow the format as specified.

• └─ Dataset/
      ├── image_00/ # aligned images from different cameras
      │   └── data/
      │       ├── <name1.ext>
      │       └── <name2.ext>
      ├── image_01/ # aligned images from different cameras
      │   └── data/
      │       ├── <name1.ext>
      │       └── <name2.ext>
      ...
      │
      └── velodyne_points/ # 3d point clouds
          └── data/
              ├── <name1.pcd>
              └── <name2.pcd>

Create a new annotation task

• Go to "Datasets", choose your Dataset.

• In "Basic", click "Create new task".

Step 1: Task info

• In Create Task, fill in the first step "Task Info": name your task then choose the dataset.

• Click "Continue".

Step 2: Select data

• Choose your data by these ways: click on the photo or choose the "Select/Deselect All"

• Filter your data by clicking "Filter", it will show image's status or "Date created".

• Choose "View Selected" if you want to check which one you have chosen for this task.

• Click "Continue".

• During the progress, you can always change the information of creating task by click "Back".

Step 3: Quality control

In this step, you will set the percentage of quality metrics base on your expected quality then click "Continue":

• Precision: shows how accurate the annotations (by the annotators) are, as compared to the reviewer(s). High precision means most annotations by the annotators are correct.

• Recall: shows how well the annotators are at identifying the objects of interest. High recall means most of the objects are correctly identified, whereas low recall means most of the objects are missing.

• Review sampling: the fraction of data to be sent to the reviewers for feedback.

Notes: These metrics are concerning the performance of the annotators.

Step 4: Assign workers

This step will allow you to assign work to your members. Each task needs at least one annotator and one QA.

Assign In house Worker

1. First, choose the source of worker by clicking "In house".

2. In "Team member", please be specific with your member's account and his role. You can also copy or delete member.

3. Also if your task need more member, please click "Add member".

4. Click "Submit".

In house members are member from your team. Remember to add them in your team first then assign task later.

Assign Outsource Worker

1. First, choose the source of worker by clicking "Outsource".

2. Fill in team member's email, phone number accurately and write down your requirement in detail.

3. Click "Submit".