Skip to content

Wake collections

Auto repair video dataset.

Review first-person repair work around tools, components and confined workspaces. Inspect the relevant contact points before defining an automotive collection.
Auto repair video dataset. Collection recording still.
Repair actions and tool use

Open the auto repair video dataset. collection

Inspect access and visibility.

A tool may enter an area the camera cannot see. Check whether the component, tool contact and resulting change are visible rather than inferred.

Keep inspection and repair distinct.

An inspection, adjustment and replacement can be different learning targets. Mark the start state, intervention and final check for the target you choose.

Define the automotive brief.

Specify the component, task, workspace and required viewpoints. Record only work the contributor is qualified to perform, under the applicable task and vehicle instructions.

Questions, answered.

What makes repair footage useful to review?

Clear views of the component, tool interaction and result let a team assess whether the footage answers its learning question.

Are every vehicle and repair procedure covered?

No coverage should be assumed from the collection name. Inspect the viewer and discuss missing tasks with Wake.

Keep exploring.

What should your robot learn next?

Bring the task, setting and signals you need. We’ll work through the collection and annotation requirements with you.

Build with Wake