Egocentric Drywall and Wall Finishing Video Data

Egocentric drywall and wall finishing video data is first-person video of a person doing drywall work, recorded from the contributor’s own point of view with synchronized sensor streams alongside. The camera sits on the head, close to what a robot’s onboard camera would see at a wall, so the frames show what the hands are doing to the panel and the joint rather than a distant view of someone standing near a job. On this site it is a custom collection: you name the task and the setting, and capture is run to a written skill spec.

The catalog describes the tasks in its own words, for example "Lift and screw drywall panels to studs, tape joints and apply joint compound." This page repeats those descriptions as written and adds nothing to them. The data is human capture: a paid, consented contributor wears the rig and performs the task. It is not robot or teleoperation data, and it is not a fixed corpus you download.

People reach this topic through several phrasings: drywall robot training data, a wall finishing video dataset, egocentric taping and sanding data, or custom construction task data collection. They describe one request from different angles. Under each of them the work is the same: write down the task, collect it on demand, and receive validated episodes with the labels a training pipeline expects. One of those phrasings needs a caution. The catalog states taping and mudding for drywall, but it states no sanding of drywall, so this page does not claim it.

Two rows of the site’s commercial task catalog describe drywall work. The first table below reproduces them exactly as the catalog states them. Both sit in the Active Construction Site environment of the Construction & Trades sector, both are rated Medium, and both list the capture setup "Phone (head-mounted) / multiview rig".

The first row, "Drywall Hanging, Taping & Mudding", is described as "Lift and screw drywall panels to studs, tape joints and apply joint compound." Its status is Ready to deploy. The second row hangs and tapes drywall panels, and is described as "Hang drywall panels with a screw gun and tape and mud the seams with a taping knife." Its status is Active collection. The catalog’s own wording for that second row names two hand tools; this page quotes the row and says no more about tools than the row does.

The site defines the statuses plainly. Ready to deploy means contributors and site access are set up and capture starts once a spec is signed. Active collection means recording is running now, so a spec can join an existing capture stream. The Active Construction Site environment holds 15 rows in all: six are rated Hard, seven Medium and two Easy, and eleven are Active collection and four are Ready to deploy. Drywall is two of them.

The catalog is a list of what is captured or available, not a ceiling on what can be briefed. A wall finishing step outside these two rows is a skill-spec question to settle in a Spec Pilot rather than an assumption. That includes sanding a taped joint, applying a further coat, patching a damaged wall and finishing a corner. The site lists none of those as drywall rows, so none is claimed here. The site also does not describe a panel size, a wall or ceiling position, a room type or a task duration for these rows.

Each attempt is delivered as one episode. For the instrumented tier the site describes a head-mounted stereo rig that records first-person RGB, with a metric depth stream recovered from the calibrated stereo rig rather than from a standalone depth sensor, and 6-DoF IMU data at 200 Hz from head- and wrist-mounted sensors. The wider stack is ten streams: a 4K head camera, two wrist cameras, depth, IMU, 3D hand and body pose, instance segmentation and action labels, hardware-aligned under 2 ms. The catalog lists one capture setup for both drywall rows, a head-mounted phone or a multiview rig, and which tier a program uses is set in the skill spec.

Hand pose is delivered for both hands as 21 joints per hand in three dimensions, with a per-joint visibility flag for occluded fingers. Drywall work puts the hands close to the surface and to each other, and a hand closing around an item hides fingers. The flag marks the joints that were not visible instead of presenting a guess as a measurement. Contact events mark when a hand meets the item.

Every stream is hardware-triggered onto one PTP clock domain, so depth, motion and hand pose line up with the same video frame. An episode is accepted only if RMS reprojection stays under 0.28 px. The only language field the site ties to Firsthand capture is one task instruction per episode. The site does not list a force or tactile stream for these rows, or a finish-quality measurement.

Inside each episode, action segments divide the timeline into 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, such as lifting the panel to the studs, taping the joint or applying compound over it, so one hanging-and-finishing job becomes a sequence of labeled steps rather than one long clip. The other standard fields are hand_joints and hand_visible for 3D hand pose, contact_events, objects, and per-frame instance masks where the spec calls for them, stored at the depth resolution of 848×480 so they line up with the depth stream.

Delivery is in the format your stack already reads: RLDS, LeRobot, WebDataset, HDF5, zarr or Rerun .rrd, or raw streams with metadata. The delivery formats table below gives the official description of the three that have an official source. In the HDF5 layout the site documents, hand pose sits at annotations/hand_joints and action segments at annotations/action_segments.

The site does not list labels for the quality of wall finishing. There is no stated label for whether a joint is smooth, whether tape is flat, whether compound is evenly applied or whether a wall is ready for the next coat. The verb and noun on each action segment is the vocabulary the skill spec defines with you, and anything beyond it is a scoping conversation, not a delivered field.

The GR00T N1 paper structures its training corpora as a data pyramid. In its words: "large quantities of web data and human videos lay the base of the pyramid; synthetic data generated with physics simulations and/or augmented by off-the-shelf neural models form the middle layer, and real-world data collected on the physical robot hardware complete the top." In that framing, first-person human video of drywall tasks would belong to the human-video base. Human capture does not replace robot data, and this page does not claim that it does.

The same paper says of the human video datasets it used: "These do not include explicit action labels but contain extensive sequences of human-object interactions." That is a statement about those datasets, not about this one. A Firsthand episode carries labeled action segments, 3D hand pose and contact events, so the structure of the step is explicit in the data. No model named on this page was trained on Firsthand data, and this page does not suggest otherwise.

Hanging, taping and mudding are a useful family to specify because they are ordered steps on one surface: position the panel, fix it, close the joint, cover it. The catalog rates both rows Medium. This page does not claim that any robot performs drywall work, that any model improves from this data, or any result. A custom collection lets you name the step your model is missing and have it captured, labeled and checked to terms you agreed.

Both drywall rows sit in the Active Construction Site environment, which lists an active construction site, carpentry, electrical installation, plumbing, drywall, tiling and painting among its examples. The site defines two capture modes. A head-mounted phone records the first-person view of real work on site. A multiview rig with handheld grippers records a staged working set. Both drywall rows list "Phone (head-mounted) / multiview rig" and do not say which applies, so the site does not settle whether a given drywall batch is real work at a real job site or a staged set. That is agreed in the spec. The site names no building, contractor or venue, and does not state job-site types, site counts, regions or cities for these rows.

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. At ingest, automated detection finds faces, plates and screens and blurs them irreversibly, and speech is scrubbed of names, addresses and card numbers. 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 a machine-checked reject criterion. The anonymization record is delivered with the data.

If your program needs terms about premises, access or who may be in frame, raise them when you scope it. The site states nothing about safety, compliance, certification or trade licensing for these rows, and this page makes no such claim. 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 state which of those applies to the drywall rows, so that is settled in your quote. The status of each row helps: the first row is Ready to deploy, so capture starts once a spec is signed, and the second is Active collection, so a spec can join an existing capture stream. After the first batch, on-demand top-ups run against the same accept criteria, and the site does not state a turnaround for those later batches.

Validation runs against signed accept criteria written into the spec before capture. An episode is rejected, among other criteria, when the manipulated object is out of frame for more than 15 percent of the episode, when more than 35 percent of depth pixels in the manipulation region are invalid, or when an undocumented bystander appears. A human reviewer then checks each episode against the task spec. A reject log ships with each batch, only footage that clears the criteria is billed as a validated hour, and rejected items are recaptured at the provider’s cost.

The license is perpetual, irrevocable and worldwide, with the data buyer-owned and non-exclusive by default. Training, evaluating, fine-tuning and deploying models, including commercial ones, is permitted, and exclusivity is available as a separate term. The executed agreement controls. 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.

Describe the drywall steps, the setting and the variation 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.

Three choices matter most for this family. Name the step and the item rather than a broad activity, using catalog wording such as "tape joints and apply joint compound", because that is what lets each step become one action segment. Name the setting and the capture tier, since both rows list a head-mounted phone or a multiview rig and the choice between them is yours to settle. And write the reject rule and the acceptance test first, so the batch is judged against terms you agreed.

Where this site does not state something, this page says "Not stated" rather than filling the gap. The final tables separate what is delivered from what is not stated. There is no stated sanding row for drywall, no finish-quality label, no force or tactile stream, no safety, compliance or licensing position, and no job-site type, material or panel size. Counts of episodes, hours, sites and contributors are not fixed numbers on this site. They are set in your skill spec.

The tradeoff to plan for is the embodiment gap between a human hand and a robot end-effector, which retargeting or pairing with robot data has to close. What you get is human first-person demonstrations in which the hands, the work and the step structure are explicit.