Polybox segmentation for clicking on screen

About OpenTrain

OpenTrain is a central job board for data-labeling and AI-training work. We aggregate projects from many AI companies and labeling platforms so you can find image, audio, text, and other annotation jobs in one place.

Creating an OpenTrain account is free and applying to projects usually takes only a few minutes.

• Centralized listings for AI training and data labeling roles
• Free account creation and quick applications

About AI Training Work

AI training (also called data labeling or annotation) is the human work that teaches models how to see, read, and act. For image tasks, people mark objects, draw polygons, and indicate acceptable interaction points for models to learn from.

This project is focused on game screenshots where annotators define the clickable area of an object so downstream systems can interpret user objectives correctly.

• Work directly contributes to how AI systems interpret images and user intents
• Typical tasks include drawing masks, bounding shapes, and labeling interaction targets

The Role

We need an annotator to create polygon masks around doors in video game screenshots. The project requires labeling roughly 1,000 images where the instruction is to 'Click on the door.'

This is an entry-level, part-time contractor role for contributors worldwide. You'll use Labelbox and produce polygon label outputs that include all acceptable click points on the door.

• Project scope: about 1,000 images (game screenshots)
• Input image example: a house with a door in a garden and a tree
• Input text for each image: 'Click on the door'
• Output required: polygon shape of the door that includes all acceptable click points
• Labeling software: Labelbox
• Label type: POLYGON

What You'll Do

Using Labelbox, draw precise polygon masks around the door in each screenshot so that the polygon covers every point where a click would reasonably fulfill the prompt 'Click on the door.'

Work carefully on object boundaries so the polygon does not miss acceptable click areas or include unrelated regions.

• Open each image and read the instruction: 'Click on the door'
• Create a polygon that covers the full clickable area of the door
• Ensure polygons include all acceptable click points and stay within the door region

Requirements

This is an entry-level task and requires no prior annotation experience. It helps if you have played some video games before so you can quickly recognize doors and interactive objects in screenshots.

The project is open worldwide and does not require specific languages or country residency.

• Experience level: Entry level
• Helpful: prior experience playing video games
• Worldwide applicants accepted
• No specific languages or country restrictions listed

Pay, Schedule, and Employment Type

Compensation is hourly at a rate of $8.00 USD per hour (PAY_PER_HOUR). The role is part-time and contractor-based with an expected commitment of less than 20 hours per week.

You will be hired as a contractor and can work remotely from anywhere unless additional project instructions say otherwise.

• Pay: $8.00 USD per hour (PAY_PER_HOUR)
• Workload: Less than 20 hours/week
• Employment types: CONTRACTOR, PART_TIME
• Open to worldwide applicants

How It Works and How To Apply

If you’re interested, create a free OpenTrain account and submit a quick application — it usually takes only a few minutes. This position is listed for OpenTrain AI and uses Labelbox to complete tasks.

Once accepted you will receive instructions and access to images and the Labelbox workspace. Follow the project’s polygon guidelines for consistent labeling.

• Create a free OpenTrain account to apply
• Project client: OpenTrain AI
• Labeling platform: Labelbox
• Expect to receive project guidelines and image batches after onboarding