Egocentric Commercial Laundry and Linen Handling Video Data

Egocentric commercial laundry and linen handling video data is first-person video of a person doing laundry work in a commercial setting, recorded from their 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 sorting cart, a washer door or a folding table, so the frames show what the hands are doing to the linen rather than a distant view of a room. The same idea sits behind any commercial laundry robot training data request, applied here as a custom collection scoped to your tasks.

On this site the data is human capture: a paid, consented contributor wears the rig and performs a task from a written skill spec. It is not robot or teleoperation data. This page is broader than the ironing page, which covers pressing and garment handling alone, and different from the household laundry folding page, which covers folding and sorting in homes. It is also different from the commercial laundry environment page, which is a catalog listing of tasks and their status rather than a description of the data you receive.

The task catalog lists a Commercial Laundry & Linen Service environment under Facilities & Cleaning. Its examples are a commercial laundromat, a hotel linen facility, a dry cleaner and a hospital linen service, covering sorting, washing, folding and pressing. It has five task rows, each rated Easy and recorded with a head-mounted phone and a multiview rig. Four are Active collection, which means recording is running now, so a spec can join an existing capture stream. One is Ready to deploy, which means contributors and site access are set up and capture starts once a spec is signed.

Hotel and Lodging adds linen rows of its own, all Ready to deploy and recorded the same way: industrial washer and dryer operation, soiled-linen sorting and loading, ironing and folding on an industrial press and folding table, linen changes, and hotel bed making. A separate Hospitality Service Set lists folding towels and linens by type as Available on request, recorded as a staged set with handheld grippers.

Which rows a program needs, and in what order, is a skill-spec decision. The site does not state equipment models, facility types or sizes, load volumes or cycle times, so this page does not either.

Each attempt is delivered as one episode. On this site the standard episode carries action segments, 3D hand pose for both hands with a per-joint visibility flag, contact events that mark when a hand meets an item, an objects list, and per-frame instance masks stored at the depth resolution of 848×480. The capture setup table below lists the streams and sync figures the site states for the instrumented tier: ten streams hardware-triggered onto one PTP clock domain and aligned under 2 ms, 6-DoF IMU data at 200 Hz, and calibration twice per session.

The catalog lists the capture mode for these commercial rows as a head-mounted phone with a multiview rig, while the ten-stream stack is the instrumented tier described on the capture page. How the two combine for a given laundry program is not stated, so it is confirmed in the skill spec. Force and tactile readings are not among the listed streams.

Action segments divide an episode into start- and end-timestamped spans, each carrying start_ns, end_ns, verb and noun. For laundry, a segment might be one sort into a cart, one washer load or one fold, as written in the skill spec. The sorting row names the sorting keys the catalog states, colour, fabric and customer, so a spec can write those into the step definitions.

Validation runs against signed accept criteria written into the spec before capture, and a reject log ships with each batch. Episodes are machine-checked against stated reject criteria, including a sync breach, calibration drift, the manipulated object out of frame for more than 15 percent of the episode, an undocumented bystander, and a mismatch against the named skill spec, and a human reviewer then checks each episode. Only footage that clears the criteria is billed as a validated hour, and rejected items are recaptured at our cost.

Delivery is in the format your stack already reads: RLDS, LeRobot, WebDataset, HDF5, zarr or Rerun .rrd, or raw streams with metadata, with readable converter source alongside. The license is perpetual, irrevocable and worldwide, buyer-owned and non-exclusive by default, with exclusivity available as a separate term. The executed agreement controls.

Fabric has no single shape. A sheet or towel folds, drapes and bunches as it is handled, and part of it is always under another part. The household laundry data on this site ships per-frame instance masks for that reason. Hands also cover the linen and each other, so hand pose carries a visibility flag and contact events mark when the hand meets the item. The second table below maps each challenge to the stream or label this site states.

Commercial work adds length and variety. A robot foundation model paper on arXiv describes a laundry folding task in which clothing items start in a randomized crumpled state in a bin, and the policy must take an item out, fold it and stack it from any configuration. The same paper describes a task in which a mobile robot takes laundry out of a dryer and places it in a hamper, says its multi-stage tasks take between 5 and 20 minutes, and reports that such tasks need fine-tuning with high-quality data. That is robot data, not Firsthand data, and Firsthand has no relationship with its authors.

The catalog describes the head-mounted phone as recording the first-person view of real work on site. The site also states that Firsthand does not capture at a customer’s own facility. Which facilities a given laundry program uses is not stated here and is agreed in the skill spec before capture begins.

Every contributor signs a written release, and each delivered episode links to it through consent artefact IDs in the batch data card. The consent chain has six checkpoints, and a break at any one discards the footage. At ingest, automated detection finds faces, plates and screens and blurs them irreversibly. 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, footage that cannot be de-identified is discarded, and the anonymization record ships with the data.

Names on linen tags, room numbers and patient identifiers are not listed as blur categories on this site, so this page does not claim they are blurred. The site states no separate release for bystanders who are not contributors, only that an undocumented bystander is a reject criterion. Who may be in frame, and how customer items are handled, is agreed in the spec. This page makes no hygiene, infection-control or throughput claim.

From signed spec to first validated batch, the published 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. Four of the five commercial laundry rows are Active collection, and the fifth is Ready to deploy, so the catalog status for this environment is the favorable end of what the site lists. Whether coverage is deep for your exact task is confirmed when the spec is quoted.

At program level, a pilot runs two to three weeks to first delivery and a standing program reaches full rate in four to six weeks. Those are program milestones that assume coverage is in place, and they sit beside the per-batch figures above rather than replacing them.

Describe the stations and their order, the linen items and their start state, what counts as a finished step, and the failure or variation cases your pipeline is missing. That becomes a written skill spec, quoted for reach, timeline and price before any capture begins. Our skill spec guide walks through how to write one.

Most programs begin with a Spec Pilot, a small, fast first batch that proves the spec and the accept criteria on real footage before volume scales. Collections are billed per validated hour. Counts of sites, contributors, episodes and hours are set in the skill spec, since the site does not state them for this environment.

Where the site is silent, this page says so. No equipment list, facility type or size, load volume, cycle time, number of sites or number of episodes is stated for commercial laundry. There is no linen-state, load-quality, fold-quality or success label, no force or tactile stream, and no robot result. The delivered-versus-not-stated table below separates what is stated from what is not.

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. Nothing here claims a training outcome.