Project Overview

You want to create a project that enables you to label parts of human faces.

To create this project, you must perform the following tasks:

1. List out your project requirements.

2. Identify sample images that you can use for labeling.

3. Configure project metadata.

4. Manage your project input and output fields.

5. Create the workflow that you want to implement in your project.

6. Add users to your project and assign them project roles.

7. Add a dataset to the project.

8. Start labeling input images.

This document explains how you can perform each of the tasks listed above. Specific sections in this document also contain sample data that you can use to easily create and implement this project in Taskmonk.

Listing Project Requirements

In this project, you want to:

• Label the following parts of the human face: Hair, Eyes, Ears, Nose, Lips, Chin.

• Provide details related to hair color, eye color, beard color, and so on.

Sample Input Data

A simple search in Wikimedia Commons yields a series of usable images. You select a few images and paste their links into a Microsoft Excel sheet under a column labeled image_URL. You save the Microsoft Excel sheet as Face_Labeling_Source.xlsx in your hard drive.

Download Source File

You can follow the steps listed above to create your Microsoft Excel sheet; you can also download and use this file in your project: Face_Labeling_Source.xlsx.

Configuring Project Metadata

Project Metadata is the first tab that appears when you create a project. The Project Metadata tab enables you to provide basic information, such as the name, process, and project type, associated with your project. You can also upload any documentation that you may want to add to your project.

1. To create the project, click the Create Project floating button on the left side of the Projects page.
   
   

   Open

   

   
   The Project Metadata tab associated with your new project appears.
   
   

   Open

   

    

2. Enter Parts of the Face as the Project Name and Image Labeling as the Process.

3. Select Annotation as the Project Type from the options provided.

4. Click Next.
   The Documents sub-tab appears.

5. You can upload documents associated with the project if required. This is an optional step, and you can skip it for now.
   Click Next.

The Task Design tab appears. Use this tab to manage your project input and output fields.

Managing Project Input and Output Fields

Taskmonk uses the project type that you specify to add input and/or output fields to projects as required. You can modify these later. In this instance, you selected Image/Video Annotation as the Project Type, and Taskmonk automatically adds the following fields and corresponding field types to the Task Design sub-tabs.

• Input Field

  • Field Name: MediaUrl. Field Type: Image

• Output Field

  • Field Name: Annotations. Field Type: Annotation

  • Field Name: Classes. Field Type: Class

Updating Input Field Details

You can update the Field Name in the Input Field tab to match the column header that you used in the input data file that you created earlier. In our instance, the input Microsoft Excel file has only one column, Image_URL.

• Click the edit icon adjacent to the default field name, MediaUrl. This makes this field editable. Update this field to read Image_URL and click the Update field name icon adjacent to it. The page reloads and displays the updated value in the Field Name column.

• By default, Taskmonk sets the field type to Image. This is appropriate for our requirement, and you need to make no changes in this instance.
  
  

   

   

Updating Output Field Details

You must now review and update the automatically added output fields.

1. Click the Output Field tab to display the Output Field UI.
   
   

   
   As discussed above, Taskmonk creates output fields when you specify the project type. We shall now update the image annotation fields created to meet our requirements.

2. You need to make no changes to the Annotations field.

3. The Classes field also requires no changes; however, the Possible Values column for the Classes field requires additional details. Creating this detail will enable you to create classes for hair, eyes, nose, lips, and chin to be labeled in your input images. Also, you want to identify hair color, eye color, and beard color, and must create specifications for these.
   
   To create classes associated with these specifications, click the Edit Classes icon under the Possible Values column in the row associated with Classes.
   The Manage Possible Values page appears.
   
   

   
   Use the Manage Possible Values page to create classes and attributes associated with the labels that you want to create.
   
   For example, let’s create a class and its attributes for ‘Hair’.
   
   

4. Enter Hair in the Class field.

5. If you plan to have multiple classes, it may be a good idea to configure hotkeys. These function as shortcuts to specific classes. In this instance, you decide to skip this step.

6. Enter the description of the class in the Class Description field. This description appears when you hover over the Head class in the labeling UI.

7. You notice that you also have an attribute, Hair Color, that must be configured. To add attributes to this class, click Add Attributes. The field-set expands to display fields associated with the attribute details.

8. Enter Hair Color as the name of the attribute in the Field Name field.

9. Click the Field Type field and select DropDown. This ensures that, when the labeling analyst clicks the Hair Color field, the options associated with the attribute shall appear as a drop-down list.

10. Enter the values that you want to associate with this attribute in the Possible Values field. In this instance, enter Black, White, Golden, Brown, NOTA.

11. Click the Default Value drop-down list. You will see that it displays all the attribute values that you specified in the Possible Values field. Select the value that must populate this field by default when the labeler selects Hair as the label class.

12. Click Update to proceed.
    The Added Classes table appears on your right-hand side with the details you created displayed.
    
    

     

13. You can similarly create additional classes and attributes for eyes, nose, ears, lips and chin.

14. Once you have created the classes that you need for your project, click Save & Close to save your changes and return to the Output Fields tab.

15. Click Next to move to the next step, Quality Workflows.

Creating Quality Workflows

The Quality Workflows tab enables you to specify how you want to ensure output quality. It also helps you create the execution levels required for your quality workflows. For example, if you want to have a QA analyst reviewing labels, you can create this role using this tab.

1. In this instance, you want to enable a Maker-Editor workflow. Click the Execution Method field and select Maker-Editor from the drop-down list that appears.

2. By default, Taskmonk creates the Analyst role for you. This is the role that performs the labelling. You only need to add the QA Analyst role. To do so, click the Add Execution Level button on the right side of the page. The Add Execution Level modal appears.
   
   

    

3. Enter QA Analyst in the Execution Level Name field.

4. You can skip all other fields here. Click Add.
   Taskmonk adds the new execution level, closes the modal, and displays the updated Quality Workflow tab.
   
   

Managing Users and Roles

You must now add users to your project and assign the execution levels you just created to them.

1. Click the Users tab just above the Quality Workflow tab.
   The Users > Manage Users tab appears.

2. Click the Add button in the top-right section of the tab.
   The Select Users tab appears.

3. Corresponding to each execution level, click the Select Users field and select the desired user from the drop-down list that appears.
   
   

    

4. Click Add to add the selected users to the project.

5. Close the modal. The Manage Users tab reloads to display the updated user details.
   
   

    

Managing Project Datasets

Your project is now configured. Congratulations!

Before you can start labeling, you must upload the image data containing human faces.

1. Click the Datasets tab. The Datasets page appears. Use this page to manage datasets for your project.

2. Taskmonk organizes datasets into batches to simplify management and tracking. To add a new dataset, click Add Batch on the right side of the page. The Add Batch modal appears.

3. Enter Batch_1 as the name for the batch that you want to import in the Add New Batch field. You can ignore the other fields.
   
   

    

4. Click Submit. This creates a new batch of data for your project and adds it to the Pending tab of the Datasets page. You can now upload datasets into the batch, as required.

5. To add a dataset to the batch, click the Import button under the Tasks (Import/Export) column.
   The Import Task modal appears.

6. Click Choose Files, select a file from your computer, and click Import.
   
   

    

7. Once the dataset is imported, click Close to exit the modal.

Labeling Images Using Taskmonk

Your project is now ready for work.

1. Log in as an analyst and click the My Tasks icon at the top of the page.
   The Tasks page appears.
   
   

    

2. Click the Get Tasks button adjacent to the Parts of the Face project.
   The labeling UI associated with this project appears.
   
   

   
   You can see the following project details in the labeling UI:

   • Batch Name (Batch_1) in the top-left section of the page.

   • Classes (Hair, Eyes, etc.) on the right side of the page.

   • Class Attributes (Hair Color) on the right side of the page, below the Classes section.
     
     For detailed information on working with a typical labeling UI, see Labeling Data.

Downloadable Sample Files