Bed making and linen changes
Requirements
Fitting a sheet, tucking a corner, placing a duvet and setting a pillow each put the linen in the hands and the bed under them. First-person capture keeps both hands and the cloth in frame, and the two wrist cameras add a closer view of the hands than the head camera alone. 3D hand pose carries a per-joint visibility flag, so a finger hidden under a sheet is flagged rather than guessed.
The site’s skill-spec guide says to name the verb and the noun rather than a broad activity. A spec for this family names the step and the item, such as strip a bed, fit a sheet, tuck a corner, place a duvet or set a pillow, and where one completed step starts and ends, so each can become its own labeled segment.
The household chores dataset page lists "Bed-making, bimanual sheet handling" as a skill and says deformable objects dominate that data, so instance masks are shipped per frame rather than propagated. A mask that follows a sheet or a duvet has to be re-drawn as the cloth moves, and the spec decides where masks are called for.
An item is rejected when the manipulated object is out of frame for too much of the episode, when depth is invalid over too much of the manipulation region, or when an undocumented bystander appears. The thresholds are agreed in the spec, so the batch is judged against terms you signed.
Definition
Egocentric bed making video data is first-person video of a person making a bed, stripping it, dressing it with fresh linen, fitting sheets, tucking corners, placing a duvet and arranging pillows, 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, so the frames show the bed, the sheet or cover in the hands, the corner being tucked and the pillow being set, rather than a distant view of someone in a room. On this site it is a custom collection: you name the task, the bedding, the setting and the capture tier, 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 covers a different subject from egocentric hospitality and housekeeping data, which spans guest-facing and room-turnover work broadly, from egocentric laundry folding data, which centres on folding clean laundry, from egocentric ironing and garment handling data, which centres on pressing and hanging, and from egocentric deformable object data, which covers soft shape-changing items in general. The subject here is the bed itself: stripping it, fitting and tucking sheets, changing a case or a cover, and arranging the result.
People reach this topic through close cousins: bed making robot training data, an egocentric bed making video dataset, linen change task data, or custom bed-making 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 bed-size or linen-type label, a wrinkle or cloth-state label, a room-state or task-completion label, or a force or tactile reading.
Why it matters
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 bed making and linen changes sits in that base layer. Human capture does not replace robot data, and this page does not claim that it does.
A bed is a hard case for reasons the site already states. The household chores dataset page lists "Bed-making, bimanual sheet handling" among its skills and describes that family as large deformable objects and two-handed coordination. The same page says deformable objects dominate that data, so instance masks are shipped per frame rather than propagated. Order matters as well. The catalog describes one row as stripping the bed, fitting a protector and sheets, remaking it and arranging the pillows, and another as fitting sheets, tucking corners, placing the duvet and arranging pillows. A clip that shows a bed somewhere in frame does not carry what a policy needs. The hands, the linen and the bed 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 bed making that points at the spread of bedding, settings and steps. A custom collection lets you aim at the step your model is missing, such as placing a duvet or changing a case, rather than hoping a broad collection happens to contain it. This page does not claim that any public corpus lacks bed-making data.
Bed-making and linen tasks
The first table below reproduces six rows of the site’s commercial task catalog for the hotel and lodging environment and the hospitality service set. The hotel and lodging environment is described with examples of hotel guest room servicing, resort housekeeping, bed-making, linen changes, bathroom cleaning and amenity restocking. Four of its rows are about the bed, all easy and all ready to deploy: "Hotel Bed Making", "Linen changes", "Strip and remake guest beds using fresh linens and mattress protectors" and "Hospitality Housekeeping", which strips and makes beds as one part of a wider room pass. The two service set rows, "Make up room bedding" and "Fold towels + linens", are medium rows captured with a multiview rig and handheld grippers and are available on request.
The second table covers the catalog’s care and health settings and the cloth bench. The care room is described with the examples "Care room: bed, tray table, linens, participant or mannequin", and its three rows are "Make a bed", which spreads, tucks and sets pillows, "Change bed linens", which strips and re-makes bed linens and is a medium row, and "Arrange comfort items", which arranges pillows and blankets. Two further rows, "Patient room prep" and "Room preparation", make the bed as one step of preparing a room. All of these are available on request. The catalog’s example line for the hospital environment carries a note on HIPAA compliance. This page makes no compliance claim, and what applies to your program is settled in the spec.
The same second table lists two cloth bench rows that isolate one hard step: "Insert duvet cover", which inserts a duvet into its cover and is a hard row, and "Flatten + smooth fabric", which flattens a sheet and smooths wrinkles. The bench rows carry a different capture setup from the hotel rows, a multiview rig with handheld grippers, and both are available on request.
One catalog row says "to hotel standard", and the site does not define that standard, so none is stated here. The site does not describe a bed size, a mattress type, a linen fabric or weight, a duvet weight or a pillow count for any row. A task outside these rows is settled in a Spec Pilot rather than assumed.
The site names no property, brand, room count or facility for any of these rows, and it lists capture at your own facility as not offered. This page therefore makes no claim that episodes are recorded in a particular hotel, hospital or care home. The environment names are the catalog’s own, and which of them applies to your program is set in the spec. Laundry-service rows such as sorting soiled linens, folding and pressing are neighbouring topics and are not covered here.
Episodes
Each attempt is delivered as one episode, and a bed-making task needs a decision up front about what one attempt is: one corner tucked, one pillow set, one duvet placed, one bed made from bare mattress to finished, or one strip-and-remake 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 linen change episode might open with stripping the used linen and close with the last pillow arranged, and a bed-making episode might run from fitting the bottom sheet to placing the duvet, as the catalog rows describe. These are illustrations of how a spec can cut an episode, not stated durations. Whether bagging used linen is part of the strip step or its own, and whether arranging pillows is a separate episode, are spec decisions.
The settings behind these rows are a hotel and lodging environment, a hospitality service set, a care room, a hospital and a nursing home environment, and a cloth bench. Which of those apply to your program is set in the spec, along with how many bedding types, bed setups and steps you want covered.
Capture setup
A bed-making 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 one hand holds a corner of the sheet and the other tucks it.
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 hotel, hospital and nursing home rows, and a multiview rig with handheld grippers for the service set, care room and cloth 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 bed making. Force and tactile readings are not among the streams, nothing measures how tight a sheet is pulled, and no robot or teleoperation data is captured.
Labels
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 linen change becomes a run of labeled steps such as strip, fit, tuck, place and arrange 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, lift and spread linen and how they move around the bed. The site describes contact events as the moment the hand meets an item, and it does not say whether a sheet, a duvet, a pillow or the mattress 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 bed or the linen itself. There is no stated label for bed size or linen type, wrinkle or cloth state, whether a bed is made to a standard, room state, or a task-completion score, and no contact force reading. Catalog descriptions such as fitting sheets and tucking corners describe the task the contributor performs, not a delivered measurement.
Stated limits
Several limits the site already states apply directly to bed-making 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, and a sheet or duvet falls over the hands constantly. Masks are the second. The household chores dataset page says clothing, bedding and bags do not track well with masks propagated frame to frame, so per-frame instance masks are shipped for them. Depth is the third. The site says depth degrades on wet and reflective surfaces and handles that with per-frame confidence maps, but it does not say how depth behaves on a particular fabric or a large flat cover.
The fourth 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. A bed is large and the contributor moves around it, so what counts as the manipulated object, the bed, the sheet or the pillow, is worth settling in the spec.
Also not stated: a mask accuracy figure for sheet edges or folds, and any description of how a tucked corner or a covered pillow is judged finished. If your policy depends on either, say so at the start.
QA, consent and safety
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.
One catalog setting needs a decision here. The care room is described as including a "participant or mannequin", and the site does not say which is used for a given row. If a person is in or beside the bed, that is a second person in the frame, and the same rule applies as to any capture: an un-consented person who is identifiable in a frame means the footage is de-identified or discarded, so settle who is in the scene, and how they are consented, in the spec.
On safety and qualification, the site states no contributor safety protocol, no hygiene or infection-control guidance, no lifting or ergonomics guidance and no care or hospitality certification for any bed-making row, and this page does not describe one. 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.
Delivery and speed
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 bed-making 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 bedding types, bed setups 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.
Not stated
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 bed size, linen type, wrinkle state or made-to-standard, no force or tactile stream, no mattress or linen specification, and no property, hotel or care-facility claim beyond the catalog rows quoted above. Nothing here claims a training outcome or a model improvement.
The tradeoff to plan for is the embodiment gap between a human pair of hands spreading and tucking linen and a robot platform doing the same, which retargeting or pairing with robot data has to close. Counts of contributors, hours and tasks are set in your skill spec.
Getting started
Describe the bedding, the steps, 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 bed-making 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 item rather than a broad activity, because that is what lets each step become one action segment. Name the setting and the capture tier, since the hotel rows and the care room rows are listed with different capture setups. And name where the episode starts and stops: one corner, one pillow, one bed from bare mattress to finished or one full strip-and-remake sequence, since the site does not set an episode length.
Capture setup
| Part | As stated on this site |
|---|---|
| Instrumented rig | Head-mounted stereo rig worn by the contributor; first-person RGB |
| Streams | Ten: 4K head camera, two wrist cameras, depth, IMU, 3D hand and body pose, instance segmentation, action labels |
| Sync | Hardware-triggered onto one PTP clock domain; aligned under 2 ms |
| Depth | Metric depth recovered from the calibrated stereo rig; per-frame confidence maps are shipped; degrades on wet and reflective surfaces |
| IMU | 6-DoF at 200 Hz from head- and wrist-mounted sensors |
| Calibration | Twice per session; accepted only under the 0.28 px RMS reprojection gate |
| Masks | Per-frame instance masks where the spec calls for them, stored at 848×480; shipped per frame, not propagated, for clothing, bedding and bags |
| Catalog capture setup | Head-mounted phone or multiview rig for the hotel, hospital and nursing home rows; multiview rig with handheld grippers for the service set, care room and cloth bench rows |
Every collection runs through the same seven-stage end-to-end custom collection pipeline.
Hotel and service set rows
| Task (as listed) | Description (as listed) | Environment | Skill level | Capture setup | Status |
|---|---|---|---|---|---|
| Hotel Bed Making | Fit sheets, tuck corners, place the duvet and arrange pillows to hotel standard. | Hotel & Lodging | Easy | Phone (head-mounted) / multiview rig | Ready to deploy |
| Linen changes | Strip used linens from the bed, bag them and dress the bed with fresh linen. | Hotel & Lodging | Easy | Phone (head-mounted) / multiview rig | Ready to deploy |
| Strip and remake guest beds using fresh linens and mattress protectors | Strip the bed, fit protector and sheets, remake it and arrange the pillows. | Hotel & Lodging | Easy | Phone (head-mounted) / multiview rig | Ready to deploy |
| Hospitality Housekeeping | Strip and make beds, clean bathrooms, dust and vacuum and restock guest rooms. | Hotel & Lodging | Easy | Phone (head-mounted) / multiview rig | Ready to deploy |
| Make up room bedding | Change case and spread cover. | Hospitality Service Set (table, bar, room) | Medium | Multiview rig + handheld grippers | Available on request |
| Fold towels + linens | Fold towels and linens; stack by type. | Hospitality Service Set (table, bar, room) | Medium | Multiview rig + handheld grippers | Available on request |
Care, health and bench rows
| Task (as listed) | Description (as listed) | Environment | Skill level | Capture setup | Status |
|---|---|---|---|---|---|
| Make a bed | Spread, tuck, and set pillows. | Care Room (assistive care) | Easy | Multiview rig + handheld grippers | Available on request |
| Change bed linens | Strip and re-make bed linens. | Care Room (assistive care) | Medium | Multiview rig + handheld grippers | Available on request |
| Arrange comfort items | Arrange pillows and blankets. | Care Room (assistive care) | Easy | Multiview rig + handheld grippers | Available on request |
| Patient room prep | Make the bed, stock supplies, position equipment and set up the room for the next patient. | Hospital & Clinical Care | Easy | Phone (head-mounted) / multiview rig | Available on request |
| Room preparation | Make the bed, tidy surfaces, set out water and personal items and prepare the room. | Nursing Home & Assisted Living | Easy | Phone (head-mounted) / multiview rig | Available on request |
| Insert duvet cover | Insert a duvet into its cover. | Deformables & Textiles Bench | Hard | Multiview rig + handheld grippers | Available on request |
| Flatten + smooth fabric | Flatten a sheet and smooth wrinkles. | Deformables & Textiles Bench | Easy | Multiview rig + handheld grippers | Available on request |
Where bed making is named
| Where | Wording on the site | What it tells you |
|---|---|---|
| Commercial task catalog: hotel and lodging | "Hotel Bed Making": "Fit sheets, tuck corners, place the duvet and arrange pillows to hotel standard."; the environment lists bed-making and linen changes among its examples | Four bed rows, easy, ready to deploy; no hotel standard defined |
| Commercial task catalog: hospitality service set | "Make up room bedding": "Change case and spread cover." | Medium row, available on request |
| Commercial task catalog: care room | "Make a bed": "Spread, tuck, and set pillows."; examples "Care room: bed, tray table, linens, participant or mannequin" | Three rows, available on request; participant or mannequin not specified per row |
| Commercial task catalog: hospital and nursing home | "Patient room prep": "Make the bed, stock supplies, position equipment and set up the room for the next patient." | The bed is one step of a wider room task; available on request |
| Commercial task catalog: cloth bench | "Insert duvet cover": "Insert a duvet into its cover." | Bench-scale rows that isolate one hard step; available on request |
| Household chores egocentric dataset page | Skill "Bed-making, bimanual sheet handling"; environments include "Bedroom"; deformable objects dominate, so masks are per frame | Bed making is a stated dataset skill at home; this page covers the custom collection |
| Egocentric hospitality and housekeeping page | Hospitality and housekeeping broadly, including room-turnover passes | The neighbour; bed making is not named there as its own task family |
| Egocentric laundry folding and ironing pages | Folding laundry; pressing, hanging and sorting garments | The neighbours; making a bed is not their subject |
| Bed size, linen or mattress specification, wrinkle state, hotel standard, property or facility | Not stated | Settled in the skill spec; not claimed here |
Streams and labels
| Stream or label | What it gives a bed-making episode | Status |
|---|---|---|
| Head and wrist cameras | First-person view of the task plus a closer view of the hands | Delivered |
| Depth with confidence maps | Bed and linen geometry, with a confidence map to mask low-confidence regions | Delivered |
| 6-DoF IMU at 200 Hz | Head and wrist motion aligned to the video frame | Delivered |
| 3D hand and body pose, with per-joint visibility flags for the hands | Both hands across the step and the body’s movement around the bed; occluded fingers flagged | Delivered |
| Contact events | The moment the hand meets an item | Delivered |
| Action segments (start_ns, end_ns, verb, noun) | One labeled span per step | Delivered |
| Per-frame instance masks (848×480) | Masks where the spec calls for them; shipped per frame for clothing, bedding and bags | Delivered where specified |
| Consent artefact IDs and anonymization record | Provenance per episode | Delivered |
| Bed size, linen type, weight or fabric | Not listed | Not stated |
| Wrinkle or cloth-state label | Not listed | Not stated |
| Made-to-standard, room-state or task-completion score | Not listed | Not stated |
| Contact force, sheet tension or tactile readings | Not among the listed streams | Not stated |
| Safety, hygiene or care certification | Not among the listed streams | Not stated |
| Robot or teleoperation data | Human first-person demonstrations only | Not captured |
| Counts of contributors, hours and tasks | Set per program in the skill spec | Not stated |
Scoping checklist
| Skill-spec part | What to name for a bed-making request | Drawn from what the site states |
|---|---|---|
| Named skill | The verb and the noun: strip a bed, fit a sheet, tuck a corner, place a duvet or set a pillow, not "bed making" | Catalog wording such as "Fit sheets, tuck corners, place the duvet and arrange pillows to hotel standard." |
| Geometry that must be visible | What must be in frame and for how long: both hands, the linen, the bed | Set in the skill spec |
| Condition budget in hours | Hours per condition cell, including bedding types, bed setups and failure cases | Set per program in the skill spec |
| Reject rule | The conditions that fail an item, written first | 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 |
| Acceptance test | One test that decides whether an attempt counts as complete | Your acceptance test is agreed in the spec |
| Setting and capture tier | Hotel, service set, care room or cloth bench, and phone or multiview rig | The catalog lists a capture setup per row |
| Episode boundary | One corner, one pillow, one bed or one strip-and-remake sequence | Set in the skill spec; the site does not state an episode length |
| Participants and lighting | Handedness, technique and pace; daylight or low light; whether anyone else is in the scene | Set per program in the skill spec |
FAQ
Sources
Tell us the bedding, the steps and the settings your model needs to see. We will quote reach, timeline and price against a written spec, with a first validated batch in a median of 48 hours where coverage is deep.