Create a project

If your task contains many objects of different types, you should break it down. For example, you can ask users to select numbers in the first project, doorways and windows in the second project, walls in the third project, and plumbing in the fourth project.

The simpler the task, the cheaper it is and the better the quality of the final result. Set the cost of labeling a single class of objects in photos at about $0.01.

Use the Object selection in an image preset. You can open this preset in the editor and add a drop-down list for labeling the selected object. See how to do this in the editor description (Dropdown list tab).

You can use JavaScript to add assignment validation depending on a checkbox. An example is provided in the “Search for information online” preset.

The number of tasks depends on how difficult and time-consuming the tasks are. Keep the size reasonably small. Large task suites are unpopular, partly because they are inconvenient for Tolokers (for example, if the internet connection is unstable).

You can create such an assignment based on the classification preset. Show a product image and ask the question: "Does the product match the query?" Add two radio buttons for responses: “Yes” and “No”.

Generate the screenshots of pages and manually label areas using the "Object selection in an image" preset.

We recommend that you break down your object selection project into three projects in Toloka:

1. Sorting images containing an object.

   1. Create a task using the “Image classification” preset.

   2. Sort the images containing the object you are looking for.

   3. Show the image to the Toloker and ask if the object is in the image. Response options: Yes/No.

2. Selecting objects in images.

   1. Select the object in the images you obtained after the previous project. You already have such a project. Run the task with manual review.

   2. Use the quality control rules: fast responses, manual review, and post-review re-assessment. Description of rules with examples.

3. Reviewing object selection assignments.

   1. Create a task using the object selection preset.

   2. Hide the editor and ask whether the object is selected correctly. Response options: Yes/No.

   3. In the input data, pass the images and coordinates of the labeled objects from the previous task.

      Now you can run the resulting pool with an overlap of 3–5 or with dynamic overlap. After that, you can aggregate the results and then upload the data for review to Project 2.

      To prevent the Tolokers who worked on the second project from doing the review, assign a skill to them. Use this skill as a filter in the pools of the third project.

Main options:

• Select an arbitrary area in the image (for example, put a square in the upper-right corner). In this case, the project instructions for reviewers should also reflect this.

• Ask the Toloker to skip the task and report it in a personal message. Messages are reviewed by the requester. If the object is truly missing, the task is deleted from the pool by resetting the overlap.

• Add an additional “No object” checkbox to the interface. Make sure that your JS checks that either the object is selected or the checkbox is enabled. In this case, add information about the checkbox value in the review task interface.

You can create a selection + drop-down list with category selection. See how to do this in the editor description (Dropdown list tab).

The coordinates are relative to the image.

Use the C shortcut for closing areas.

You can also use the library to customize the keyboard shortcuts for your tasks.

To create a shortcut, add the following action to the "onKey" method:

onKey: function(key) {
    var el = this.getDOMElement().querySelector(".image-annotation-editor__shape-polygon");

    if (key === 'D') {
      el.click();
      el.classList.add('image-annotation-editor__shape_active')
    }

In the case of crowdsourcing, it's better to break down this task. The simpler the task, the cheaper it is and the better the quality of the final result. The cost of labeling a single class of objects in photos might be about $0.01.

​Base your task on the “Object selection in an image” preset. See the step-by-step guide for creating this type of project on this page.

The editor used in the preset lets you add a drop-down list for labeling the selected object. See how to do this in the editor description (Dropdown list tab).

In the standard preset with an area selection editor, you can't use the control tasks, because in order for the assignment to be accepted by the system as correct, the object selected by the Toloker must exactly match the control object. This is almost impossible. Therefore, you can leave the GOLDEN field empty in the task file or simply delete all the columns except INPUT.

You can't use training and the main pool with the Training type in an area selection project because for the response to be correct, the object selected by the Toloker must exactly match the control object. This is almost impossible.

Such tasks are usually run with manual review: the Toloker submits an assignment, and then the assignment is rejected or accepted after the review.

For pre-selection of users, you can use “examination tasks”. Review the assignments and assign skills based on the percentage of accepted assignments. For this purpose, add the “Results of assignment review” rule to the pool. To make sure that only the good Tolokers are admitted to the main pool, put a skill-based filter to the pool.

1. Use the Image comparison (Side-by-side) preset.

2. In the TSV file, specify the links to the compared images.

3. In the file, create the tasks where all the images will be compared in pairs:

   • Image 1 and Image 2

   • Image 1 and Image 3

   • Image 2 and Image 3

4. Process the results.

You can also edit the project so that the Toloker sees 3 images at once and selects one of them.

You can structure your text classification task using the source text and radio buttons. It can also be text and checkboxes if you use multiple subjects.

We recommend that you base it on the "Text classification" preset.

You can also create your own interface. All the available tools are described in the documentation.

You can see how it's implemented in the “Transcript of audio recordings” preset where a string-type output field is used. In the “Text recognition from an image” preset, you can view how to describe text input fields. Please note that if you later want to use the Dawid-Skene aggregation method, you must specify allowed values.

1. Create a project from an empty preset.

2. Write your question in the HTML block.

3. If you need an extensive response, add the required number of text entry components.

4. If you need a short response, add the required number of string entry components.

5. Come up with a name for each of the components and create a matching number of string-type output fields with the same names. They will be used to save responses.

6. Make all the output fields mandatory.

7. Make the input field auxiliary. It will only be used to create a file with tasks.

8. Come up with a name for the input field and set the string type for it (see the step-by-step guide on creating a survey).

The survey includes an auxiliary input field. You can use it to pass any information, and it won't be visible to Tolokers.

Upload one task and use the overlap parameter to enter the number of people you want to survey.