Egocentric Masonry and Plastering Task Data

Egocentric masonry and plastering task data is first-person video of a person doing hand-tool trade work, recorded from the contributor’s own point of view with synchronized sensor streams. The head-mounted camera sits close to what a robot’s onboard camera would see, so frames show what hands and tools are doing to the material. On this site it is a custom collection: you name the task and setting, and capture runs to a written skill spec.

One caveat first, because the title is a search phrase and not a catalog heading. The task catalog has no row for bricklaying, blockwork, mortar, plastering or rendering. It has one row that states concrete work. This page covers that row as written and adds nothing to it. The data is human capture by paid, consented contributors, not robot or teleoperation data, and not a fixed corpus you download.

Phrasings such as masonry robot training data, a plastering video dataset, egocentric bricklayer data or custom construction task data collection all lead to the same method: write the task down, collect it on demand, and receive validated episodes. Whether a bricklaying or plastering task can be briefed is a scoping question for a Spec Pilot, not an existing offering.

Two catalog rows mention concrete, reproduced exactly in the first table below. "Concrete work" sits in the Heavy Industry & Fabrication environment (Manufacturing & Workshops sector): "Mix, pour, screed and finish concrete into forms using shovels, floats and trowels." It is rated Hard, lists "Phone (head-mounted) / multiview rig", and is Available on request.

"Construction General Skilled Trades" sits in Active Construction Site (Construction & Trades): "Skilled trade work on active sites including framing, concrete and structural tasks." It is rated Hard and Ready to deploy. It is a general trades row naming concrete among other work, not a masonry or plastering row, so this page uses it only as context and makes no structural claim.

Statuses are defined plainly. Available on request means scoped and recruited to spec when a buyer asks. Ready to deploy means contributors and site access are set up and capture starts once a spec is signed. Heavy Industry & Fabrication holds five rows, all Hard and all Available on request: concrete work, glass manufacturing, metal fabrication, operating heavy machinery controls and welding.

The catalog is not a ceiling on what can be briefed; a step outside it is a skill-spec question for a Spec Pilot. The site does not describe a pour size, form type, surface, setting or task duration for the concrete row.

We checked the catalog for bricklaying, blockwork, mortar, plastering, rendering and stucco; none appears in a task name, description or example. The only mentions of stone are a sculpting row and a jewellery stone-setting row, which are different tasks. The examples for Active Construction Site (carpentry, electrical installation, plumbing, drywall, tiling, painting) and Heavy Industry & Fabrication (metal fabrication, woodworking, glass manufacturing, a concrete plant, a welding workshop) include neither trade.

Drywall and tiling have their own pages; they are different tasks and this page does not stand in for them. A buyer needing bricklaying or plastering would describe the steps, tools and setting, and the quote would state whether and how it can be captured.

Each attempt is one episode. For the instrumented tier the site describes a head-mounted stereo rig recording first-person RGB, metric depth recovered from the calibrated stereo rig, 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. Which tier a program uses is set in the skill spec.

Hand pose covers both hands as 21 joints each in 3D, with a per-joint visibility flag for occluded fingers; a closed grip on a float handle hides some. The flag marks what was not visible instead of presenting a guess as a measurement. Contact events mark when a hand meets the item.

Streams share one PTP clock domain, and an episode is accepted only if RMS reprojection stays under 0.28 px. The site lists no force or tactile stream and no measurement of the finished surface for these rows.

Action segments divide each episode into spans carrying start_ns, end_ns, verb and noun. Here a segment can be one step as written in the spec, such as mixing, pouring, screeding or finishing, so one concrete job becomes a sequence of labeled steps. Other fields are hand_joints, hand_visible, contact_events, objects and, where the spec calls for them, per-frame instance masks at 848×480 to match the depth stream.

Delivery is in RLDS, LeRobot, WebDataset, HDF5, zarr or Rerun .rrd, or raw streams with metadata; official repositories are linked in the sources. The site lists no label for finish quality: nothing states whether a surface is level, smooth or cured.

The GR00T N1 paper describes 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." First-person human video of concrete work would belong to the human-video base. Human capture does not replace robot data.

Of the human video datasets it used, the paper says: "These do not include explicit action labels but contain extensive sequences of human-object interactions." That describes those datasets, not this one: a Firsthand episode carries labeled action segments, hand pose and contact events. No model named here was trained on Firsthand data, and this page claims no robot performs concrete work and no model improves from this data.

The concrete row sits in Heavy Industry & Fabrication, whose examples include a concrete plant; the general trades row sits in Active Construction Site. Both list a head-mounted phone, which records the first-person view of real work on site, or a multiview rig, which adds synchronized third-person angles when the spec asks. The site does not say whether a concrete batch is captured at a plant, a construction site or a workshop, so that is agreed in the spec. It names no plant, contractor or venue.

Every contributor signs a written release, and each episode links to that record through consent artefact IDs in the batch data card. At ingest, faces, plates and screens are detected and blurred irreversibly, and speech is scrubbed of names, addresses and card numbers. A human reviewer confirms no un-consented person is identifiable, and footage that cannot be de-identified is discarded. The anonymization record ships with the data.

The site states nothing about safety, compliance, certification, building codes or trade licensing for these rows, and this page makes no such claim. 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. It does not say which applies to the concrete row, which is Available on request, so the timeline is settled in your quote. On-demand top-ups run against the same accept criteria; no turnaround is stated for them.

Validation runs against signed accept criteria. Among others, an episode is rejected when the manipulated object is out of frame for more than 15 percent of it, 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 spec. A reject log ships with each batch, only cleared footage is billed as a validated hour, and rejects are recaptured at the provider’s cost.

The license is perpetual, irrevocable and worldwide, buyer-owned and non-exclusive by default, with exclusivity as a separate term. The site publishes no price list, minimum order or batch size for this task family.

Describe the steps, the setting and the variation cases your pipeline is missing. That becomes a written skill spec, quoted for reach, timeline and price before capture, and most programs run a Spec Pilot first: a small, fast batch that proves the spec and accept criteria on real footage. A published sample recording is on the sample data outputs page.

Name the step and item using catalog wording such as "screed and finish concrete into forms", name the setting and capture tier, and write the reject rule and acceptance test first.

Where this site does not state something, this page says "Not stated" rather than filling the gap; the final table lists the gaps. Counts of episodes, hours, sites and contributors 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.