Egocentric Classroom Setup and Cleanup Data

Egocentric classroom setup and cleanup data is first-person video of one person getting a room ready for the day, recorded from that person’s own point of view while they work. A paid, consented contributor wears the capture rig and performs the task written in a skill spec, and each attempt is delivered as an episode with synchronized streams, hand pose and labeled action segments. It is human capture. It is not robot or teleoperation data, and it is not a fixed corpus you download.

The site states the task once, in its commercial task catalog. Under the Childcare & Early Education environment, the row "Classroom setup" reads: "Arrange tables, chairs and learning materials and set up activity stations for the day." It is rated Easy, captured with a head-mounted phone or a multiview rig, and listed as available on request. The site does not list a separate row for end-of-day cleanup, so this page covers setup as the catalog states it and marks cleanup as not separately stated.

Neighbouring pages cover related but different ground. Egocentric event venue setup data covers venue setup for events rather than daily classroom use, egocentric object rearrangement data is the general tabletop and household rearrangement task family, and egocentric office task data covers desk and office work rather than a classroom. The subject here is the one named catalog row: furniture, materials and activity stations in a room used for childcare or early education.

People reach this topic through classroom setup video data, school facility robot training data, egocentric room reset data, or custom facility task data collection. They describe one request from different angles. The work is the same: write the task down, collect it on demand, and receive validated episodes.

The table below reproduces the one catalog row that names this task: "Classroom setup," rated Easy, captured with a head-mounted phone or a multiview rig, and listed as available on request under Childcare & Early Education.

The row names three steps in sequence: arranging tables and chairs, setting out learning materials, and setting up activity stations for the day. The site does not break these into separate rows, so a spec that wants one of them captured on its own, rather than as part of the full sequence, needs to say that explicitly.

The closest put-away task is a neighbouring row in the same environment, "Toy and material organization": "Sort toys and materials into labelled bins and shelves after activities." The catalog also lists "Feeding" and "Low-angle child interactions," which are different tasks and are not reproduced on this page. The site states no classroom-cleanup, end-of-day reset, desk-wiping or whiteboard-wiping row anywhere in the catalog. If a program needs a reset task beyond the sorting row, it is a new task for the spec to name, not an extension of the existing "Classroom setup" row.

The site does not state a room type, furniture type, material type, or how many tables, chairs or stations one episode covers. If any of those matter for your use, raise them in the request and settle them in a Spec Pilot.

Each attempt is delivered as one episode, and classroom setup needs a decision up front about what one attempt is. It could be one piece of furniture moved into place, one activity station built, or the whole room readied from empty to set. The site does not set an episode length or the number of episodes, so the skill spec does. Inside the episode, the synchronized streams run from start to end, and action segments divide it into labeled spans.

The choice matters because of two machine-checked rules. An episode is rejected if the manipulated object is out of frame for more than 15 percent of it, and if more than 35 percent of depth pixels in the manipulation region are invalid. In a room-reset task the manipulated object changes several times, from a chair to a stack of materials to a station prop, so the spec should say which object counts as manipulated at each step and where the manipulation region is.

The site states that depth degrades on wet and reflective surfaces, and that a confidence map is delivered so low-confidence regions can be masked. It does not say how a wide room, a cluttered shelf or small, stacked materials affect depth quality for this task. If your policy depends on depth across a full room rather than a close workbench, raise it at the start and settle it in a Spec Pilot rather than assuming it.

Inside each episode, action segments divide the recording into start- and end-timestamped spans, each carrying start_ns, end_ns, verb and noun. For this task family a segment can be one step as written in the skill spec, so a room-reset run becomes a sequence of labeled steps such as moving a table, placing chairs, laying out materials and building a station, rather than one long clip. The other standard fields are hand_joints and hand_visible for 3D hand pose, contact_events for the moment the hand meets an object, objects, and instance_masks, stored at the depth resolution of 848×480 so they line up with the depth stream.

Hand pose is delivered for both hands with a per-joint visibility flag for occluded fingers. That is the label that carries how the hands grip a chair back, carry a stack of materials or place a station prop. The site describes contact events as the moment the hand meets an item, and does not say whether a table edge or a stacked material counts as the item, so the spec settles what a contact event means for your task.

The site lists no label for furniture type, material type, station layout or completion state. There is no stated room size, item count or pass-or-fail score, and no contact force or tactile reading. Camera pose, a head trajectory and a map of the room are likewise not among the listed deliverables. The site states 6-DoF IMU data at 200 Hz from head- and wrist-mounted sensors, and nothing more about motion.

Delivery is in the format your stack already reads: RLDS, LeRobot, WebDataset, HDF5, zarr or Rerun .rrd, or raw streams with metadata. Readable converter source ships alongside the data so the schema can be retargeted to your own loader. The official documentation for three of these formats is linked in the sources below. Collections are billed per validated hour, and the license is perpetual, irrevocable and worldwide, buyer-owned and non-exclusive by default, with exclusivity available as a separate term.

A published robot foundation model paper, GR00T N1, linked in the sources below, describes a data pyramid with web data and human video at the base, synthetic data in the middle and real robot trajectories at the top. First-person human video of room-reset work sits in that base layer. Human capture does not replace robot data, and this page does not claim that it does. The glossary entry on the data pyramid explains how the layers trade volume for embodiment closeness.

Room setup moves between large furniture and small materials in the same sequence. The site’s tool use page observes that grip changes between hands live in finger-level detail that RGB video alone cannot resolve, and that each change needs 3D hand pose tracked through reach, grasp and tool manipulation. That page is about driving and fastening tools, not classroom furniture, so this page does not claim the same findings for room-reset work. The reasoning carries over: a clip that shows a room somewhere in frame does not carry what a policy needs. The furniture, the materials, the hands and the step boundaries have to be in the episode, and a spec that names them lets each step become one labeled segment.

A custom collection lets a program aim at the step a model is missing, such as building one activity station or carrying a stack of materials across a room, rather than capturing the whole day at once.

This page does not claim that any public corpus lacks room-reset data, and it does not claim a training outcome, a success rate or an accuracy gain from any dataset. The tradeoff to plan for is the embodiment gap between a human arranging a room and a robot platform doing the same, which retargeting or pairing with robot data has to close.

The site describes the Childcare & Early Education setting as using a head-mounted phone or a multiview rig. It describes that setup as a head-mounted phone recording the first-person view of real work on site, with a multiview rig adding synchronized third-person angles when the spec asks for them. The site does not say more about the setting beyond the environment name.

The site names no school, district, university, employer or city, and it lists capture at your own facility as not offered. This page therefore makes no claim that episodes are recorded in a particular school or facility, and none that they are recorded at yours. Capture is run to your skill spec using the settings and contributors Firsthand arranges.

Every contributor signs a written release before capture, and each delivered episode links to that record through consent artefact IDs in the batch data card. The consent chain has six checkpoints, from before capture to after delivery, and a break at any one discards the footage rather than shipping it at a lower grade. At ingest, automated detection finds faces, plates and screens and blurs them irreversibly, and speech is scrubbed of names, addresses and card numbers while ambient audio is kept. A human reviewer confirms that no un-consented person is identifiable in any frame, and footage that cannot be de-identified is discarded. An undocumented bystander is also one of the machine-checked reject criteria.

The Childcare & Early Education environment’s own examples describe toy and material organization, feeding and low-angle child interactions alongside classroom setup. This page covers only the classroom setup row and makes no claim about capturing children, students, minors or any other person’s likeness; the privacy process above, including irreversible blurring and discard of footage that cannot be de-identified, applies to whoever performs the contracted task. None of this is legal advice, and the executed agreement controls.

From signed spec to first validated batch, the site states a median of 48 hours where coverage is deep and three to four weeks where it is thin and contributors need to be recruited first. The site does not say whether the Childcare & Early Education environment counts as deep or thin, so ask which timeline applies when you request the quote. Speed comes from a warm pipeline, not from cutting consent, review or anonymization.

The site also states program milestones: a pilot reaches its first delivery in two to three weeks, and a standing program reaches full rate in four to six weeks. After the first batch, on-demand top-ups run against the same accept criteria, so you can add steps or layouts without rewriting the spec. The site does not state a turnaround for those later batches, so plan them in your quote.

Only footage that clears the signed accept criteria is billed as a validated hour, and rejected items are recaptured at our cost. The site does not publish a price list, a minimum order or a batch size for this task family, so those are set in your quote. Collection is fast and on demand in the sense that you do not need a standing program: a focused brief is a normal starting point.

Where this site does not state something, this page says "Not stated" rather than filling the gap. There is no listed row for end-of-day cleanup, room reset, desk-wiping or whiteboard-wiping; the catalog names "Classroom setup" and the neighbouring sorting row "Toy and material organization." There is no room type, furniture type, material type or station layout beyond "tables, chairs and learning materials" and "activity stations," no task duration, no site count, no region or city, no episode count and no success or quality label. Nothing here reports a robot result, a benchmark or a model improvement.

Counts of contributors, hours and tasks are set in your skill spec. A task outside the row quoted above, including any cleanup step, is settled in a Spec Pilot rather than assumed.

Describe the furniture, the materials, the station layout and the variation or failure cases your pipeline is missing. That becomes a written skill spec, quoted for reach, timeline and price before any capture begins, and most programs run a Spec Pilot first, a small, fast first batch that proves the spec and the accept criteria on real footage before volume scales. A published sample recording is on the sample data outputs page if you want to see the streams before you write a spec.

Two choices matter most for this family. Name the step and the item rather than a broad activity such as "classroom work," because that is what lets each step become one action segment. And if the program needs a cleanup or reset task rather than setup, say so explicitly, since the catalog does not currently list one.