Egocentric Painting and Coating Video Data

Egocentric painting and coating video data is first-person video of a person applying paint, primer, stain or another finish to a surface, 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 while it works a wall, a panel or a canvas, so the frames show the roller or brush, the surface in front of it, the hands on the handle and the edge being cut in, rather than a distant view of someone decorating a room. On this site it is a custom collection: you name the task, the tool, the surface and the setting, and capture is run to a written skill spec.

The data is human capture. A paid, consented contributor wears the rig and does the task. It is not robot or teleoperation data, and it is not a fixed corpus you download. This page is a different subject from egocentric tool use data, which centres on hand tools such as drivers, clamps and measuring tools, and from wiping and surface cleaning data, which centres on cleaning a surface by hand. The subject here is applying a material to a surface with a brush, a roller or a rag, and the prep work that sits beside it. It covers what the site states about capturing, labeling and checking those tasks.

People reach this topic through close cousins: painting robot training data, an egocentric painting video dataset, brush and roller task data, or custom painting data collection. They describe one request from different angles, and the work is the same: write down the task, collect it on demand, and receive validated episodes. Nothing on this page states that the video carries a paint-type label, a coverage map or a film-thickness measurement.

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 painting sits in that base layer. Human capture does not replace robot data, and this page does not claim that it does.

A painting task is a sequence of small, ordered moves with a tool loaded with material. The catalog describes one row as masking edges, rolling primer and paint onto walls and cutting in corners with a brush, and another as rolling primer, then paint, and brushing cut-in at edges and trim. The order matters, and so do the hands: the site’s tool use page says that grip changes live in finger-level detail that RGB video alone cannot resolve. A clip that shows a roller somewhere near a wall does not carry what a policy needs. The hands, the tool and the surface have to be present in the episode, and the episode has to say where one step ends and the next begins.

Breadth matters as much as volume. A published model write-up, GEN-0, also linked in the sources, puts it this way: "data quality and diversity matters more than sheer volume." For painting that points at the spread of tools, surfaces and steps. A custom collection lets you aim at the step your model is missing, such as cutting in along trim or masking an edge, rather than hoping a broad collection happens to contain it. This page does not claim that any public corpus lacks painting data.

The first table below reproduces seven rows of the site’s commercial task catalog that name painting, priming, staining or finishing. Each is listed with an environment, a skill level, a capture setup and a status exactly as the catalog states them. It is a selection of the rows that apply a material to a surface, not the whole catalog.

Wall painting appears in two settings. The active construction site environment lists "Painting & Priming", which masks edges, rolls primer and paint onto walls and cuts in corners with a brush, and "Prime and paint interior walls using rollers and brushes", which rolls primer then paint and brushes cut-in at edges and trim. Both are medium skill with a head-mounted phone or multiview rig, marked ready to deploy and active collection respectively. That environment’s own description also lists painting among its trades. The trades workshop bench lists "Roll + cut-in paint", which rolls a panel and cuts in edges. It is an easy row captured with a multiview rig and handheld grippers, marked available on request.

Wood finishing appears in the carpentry and furniture workshop. "Staining & Finishing" applies stain, sealer or lacquer with rags, brushes or a spray gun and sands between coats, and a second row sands surfaces with an orbital sander through the grits and wipes on stain with rags. Both are medium skill, the first ready to deploy and the second active collection. The spray gun appears in that one row’s wording as a tool. The site states no spray booth, no industrial finishing line and no coating line anywhere, so this page makes no such claim.

Artistic painting appears in the art studio environment as "Canvas painting", which mixes paint on a palette and applies it to canvas with brushes and knives, and "Mural work", which sketches the outline on a wall and paints large areas and details with rollers and brushes. Both are medium skill and marked available on request. These are catalog listings, and the page does not claim more about them than the catalog does.

The second table lists neighbouring prep rows: sanding and surface prep, sanding and wiping a bench surface, caulking, and drywall taping and mudding. In a real job they sit before painting, but they are not painting rows, so they are kept apart. The nail treatments row is left out because painting nails is a beauty task, the car polishing rows are left out because polishing compound is not a paint or a coating, and the floor stripping and waxing rows are left out because they belong with floor care. The third table shows where else painting is named on the site, and where it is not. The site does not describe a paint type, a finish, a brand, a tool model or a surface size for any row, so none is stated here. A task outside these rows is settled in a Spec Pilot rather than assumed.

Each attempt is delivered as one episode, and a painting task needs a decision up front about what one attempt is: one wall, one panel, one coat, one cut-in pass along a stretch of trim, or one masking-and-rolling sequence. The site does not set that length, so the skill spec does. Inside the episode, the synchronized streams run from start to end, and action segments divide it into labeled spans.

A priming-and-painting episode might open with masking an edge and close with the last corner cut in, and a staining episode might run from the first wipe to the sanding between coats, as the catalog rows describe. These are illustrations of how a spec can cut an episode, not stated durations. Whether a coat is one episode or several, and whether sanding between coats sits inside the episode or is its own, are spec decisions.

The environments behind these rows are an active construction site, a trades workshop bench, a carpentry and furniture workshop with finishing stations, and an art studio. Which of those apply to your program is set in the spec, along with how many surfaces, tools and steps you want covered. The site lists capture at your own facility as not offered.

A painting collection uses the same capture stack as every Firsthand program. The instrumented tier is 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 two wrist cameras give a closer view of the hands than the head camera alone, which matters where a hand grips a brush or steadies a roller.

Every stream is hardware-triggered onto one PTP clock domain, so depth, motion and hand pose line up with the same video frame rather than being approximated afterwards. The rig is calibrated twice per session, before capture starts and after it ends, and an episode is accepted only if RMS reprojection stays under 0.28 px. The catalog lists the capture setup per task: a head-mounted phone or multiview rig for the construction, carpentry and art rows, and a multiview rig with handheld grippers for the trades workshop bench rows. The site does not explain the grippers further. Which tier applies to your program is set in the skill spec.

The site does not state a camera count, resolution or frame rate specific to painting. Force and tactile readings are not among the streams, no paint-type, wet-film or coverage sensing is among the ten, and no robot or teleoperation data is captured.

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 the construction row becomes a run of labeled steps such as mask, roll and cut in rather than one long clip, and a staining row can separate wiping on stain from sanding between coats. 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 hand sits on a brush or roller handle and how it moves through a pass. The site describes contact events as the moment the hand meets an item, and it does not say whether a roller, a brush, a paint tin, a masked edge or the wall counts as the item, so the spec is the place to settle what a contact event means for your task.

The site does not list labels for the paint or the finish itself. There is no stated label for paint or finish type, a coverage or painted-area map, a coat count, a wet-film or coating thickness, a stroke direction, a surface-prep state, or a task-completion score, and no contact force or pressure reading. Catalog descriptions such as rolling primer then paint or sanding between coats describe the task the contributor performs, not a delivered measurement.

Several limits the site already states apply directly to painting capture, and several it leaves open. Each stated limit below appears elsewhere on the site rather than being introduced here.

Hand occlusion is the first. Fingers disappear behind objects in manipulation, which is why 3D hand pose carries a per-joint visibility flag. Depth is the second. The site says depth degrades on wet and reflective surfaces and handles that with per-frame confidence maps. Fresh paint, stain and lacquer leave wet, glossy surfaces, but the site does not say how depth behaves on one in particular. A related statement exists for under-vehicle capture, where the site says painted and oily surfaces absorb a depth projector pattern and those episodes carry a depth-quality flag. That statement is about under-vehicle work, and this page does not extend it to walls or panels or promise a figure.

The third is the set of machine-checked reject criteria. An episode is rejected for a sync breach beyond the ceiling for more than three consecutive frames, calibration drift beyond tolerance, the manipulated object out of frame for more than 15 percent of the episode, more than 35 percent invalid depth pixels in the manipulation region, an undocumented bystander, or a mismatch against the named skill spec. A human reviewer then checks each episode against the task spec. For a painting task, the spec decides what the manipulated object is: the tool, the patch of wall or the panel being worked.

The tool use page adds that metric depth on sub-centimetre parts is flagged rather than interpolated. Whether a thin cut-in line or a bristle tip counts as such a part is not stated. Also not stated: a mask accuracy figure for brushes, rollers, tape lines or painted edges, and any description of how masking tape, drop cloths or a wet edge are handled. If your policy depends on any of those, say so at the start.

Validation runs against signed accept criteria written into the spec before capture. Each batch goes through per-batch QA and a reject log ships with it. Only footage that clears the criteria is billed as a validated hour, and rejected items are recaptured at our cost. Many teams start with a Spec Pilot, a small, fast first batch that proves the spec and the accept criteria on real footage before volume scales.

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. The anonymization record is delivered with the data.

Construction sites and shared workshops can have other people nearby. The same rule applies to them as to any capture: an un-consented person who is identifiable in a frame means the footage is de-identified or discarded, so settle bystander handling in the spec for the setting you choose.

On safety and qualification, the site states no contributor safety protocol, no guidance on ventilation, fumes, solvents, personal protective equipment or paint chemistry, and no safety certification for any painting row, and this page does not describe one. The site lists capture at your own facility as not offered. If your program needs terms about who performs the work, the materials used or the conditions under which it is done, raise them when you scope it. None of this is legal advice, and the executed agreement controls.

From signed spec to first validated batch, the timeline is 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. Speed comes from a warm pipeline, not from cutting consent, review or anonymization. The site does not say whether the painting settings count as deep or thin coverage, so ask which timeline applies to the setting you want when you request the quote. After the first batch, on-demand top-ups run against the same accept criteria, so you can add more surfaces, tools or steps without rewriting the spec. The site does not state a turnaround for those later batches, so plan them in your quote.

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. Collections are billed per validated hour.

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.

Where this site does not state something, this page says "not stated" rather than filling the gap. The table below separates what is delivered from what is not stated. There is no listed label for paint or finish type, coverage, coat count or thickness, no force or tactile stream, no paint, tool or surface specification, and no wall or panel size. There is no spray booth or industrial finishing claim, and nothing here claims a training outcome or a model improvement.

The tradeoff to plan for is the embodiment gap between a human holding a brush or roller and a robot platform working a surface, which retargeting or pairing with robot data has to close. Counts of contributors, hours and tasks are set in your skill spec.

Describe the tools, the surfaces, the setting 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. The scoping checklist below maps the parts of a skill spec onto a painting request. 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 tool rather than a broad activity, because that is what lets each step become one action segment. Name the setting and the capture tier, since a construction site, a workshop bench and an art studio are listed with different capture setups for some rows. And name where the episode starts and stops: one coat, one wall, one cut-in pass or one full masking-to-finish sequence, since the site does not set an episode length.