Egocentric Embroidery and Hand Stitching Data

Egocentric embroidery and hand stitching data is first-person video of one person stitching fabric by hand, recorded from that person’s own point of view while they work. A paid, consented contributor wears the capture rig and performs the task written in a skill spec, and each attempt is delivered as an episode with synchronized streams, hand pose and labeled action segments. It is human capture. It is not robot or teleoperation data, and it is not a fixed corpus you download.

The site states the tasks in its commercial task catalog. In the Textile Arts & Stitching Studio environment, the row "Hand embroidery" reads: "Stitch decorative patterns into fabric by hand with needle and embroidery thread." It is rated Medium, captured with a head-mounted phone or a multiview rig, and listed as ready to deploy. The row "Needlepoint" reads: "Stitch yarn through canvas in a counted pattern with a tapestry needle." It is rated Medium, with the same capture setup and status. The environment is described with examples of hand embroidery, tailoring studio, quilting, needlepoint, custom garment fitting and hemming. This page is built on those rows and on nothing else the site says about stitching.

Neighbouring pages cover different subjects. Egocentric sewing and textile handling data spans cutting, pinning, hemming, stitching and handling fabric across several environments, and it reproduces every row of the stitching studio. Egocentric garment production data centres on factory-style apparel work, and egocentric arts and crafts data covers painting, printmaking and sculpting. The subject here is the hand-stitching slice of that cluster: embroidery and needlepoint first, with the other hand-stitching rows alongside.

People reach this topic through embroidery video data, hand stitching robot training data, egocentric needlework data, or custom textile craft data collection. They describe one request from different angles. The work is the same: write the task down, collect it on demand, and receive validated episodes.

The first table below reproduces the seven catalog rows that name stitching by hand or with a needle. Five come from the Textile Arts & Stitching Studio: "Hand embroidery", "Needlepoint", "Embroidery", "Hemming" and "Tailoring". One comes from the Leathercraft Workshop, "Hand-stitching leather goods", which punches stitch holes and saddle-stitches leather pieces with two needles and waxed thread. One comes from the Deformables & Textiles Bench, "Thread lace eyelets", which threads a lace through eyelets.

Only two of the seven are hand-only fabric stitching in the studio: hand embroidery and needlepoint. "Embroidery" is described as machine and hand embroidery production of garments and textile goods, "Hemming" as measuring, pinning, pressing and stitching hems on trousers and skirts by hand or machine, and "Tailoring" as cutting, fitting and sewing garments to measure with machine and hand finishing. Those three mix hand and machine work, so a spec that wants only the hand part should say so.

The studio also lists crochet, knitting, quilting and custom garment fitting. They are textile craft rows, but they are not stitching by hand through cloth in the same sense, and the egocentric sewing and textile handling page reproduces them. This page leaves them out.

The site does not say what a stitch type, an embroidery style, a pattern or a finished piece looks like for any of these rows. It states no fabric type, thread type, hoop or frame, stitch count or piece size. Whether one episode is a single stitch, a row of stitches or a whole motif is a spec decision.

Each attempt is delivered as one episode, and stitching needs a decision up front about what one attempt is. It could be one stitch, one line of stitching, one motif or the whole piece. The site does not set an episode length, the size of the piece or the number of episodes, so the skill spec does. Inside the episode, the synchronized streams run from start to end, and action segments divide it into labeled spans.

The choice matters because of two machine-checked rules. An episode is rejected if the manipulated object is out of frame for more than 15 percent of it, and if more than 35 percent of depth pixels in the manipulation region are invalid. In embroidery the fabric, the needle and both hands sit close together, so the spec should say what counts as the manipulated object, whether that is the cloth, the thread or the needle, and where the manipulation region is at each step.

The site does not describe how fabric held in a hoop or frame, a loose thread or a hand covering the stitch is handled. If your policy depends on any of those, say so at the start, and settle it in a Spec Pilot rather than assuming it.

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 stitching session becomes a run of labeled steps 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 fingers pinch, hold and guide the needle and the cloth, and it is the main signal for the manipulation inside a stitching task. The site describes contact events as the moment the hand meets an item, and does not say whether a needle, a thread or a fold of cloth counts as the item, so the spec settles what a contact event means for your task.

The site lists no label for the stitching itself. There is no stated stitch type or count, thread type, fabric type, stitch length, seam or pattern quality, or task-completion score, and no contact force or thread tension reading. Camera pose, a head trajectory and a map of the workspace are likewise not among the listed deliverables. The site states 6-DoF IMU data at 200 Hz from head- and wrist-mounted sensors, and nothing more about motion.

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. The official documentation for three of these formats is linked in the sources below. Collections are billed per validated hour, and the license is perpetual, irrevocable and worldwide, buyer-owned and non-exclusive by default, with exclusivity available as a separate term.

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 hand work sits in that base layer. Human capture does not replace robot data, and this page does not claim that it does. The glossary entry on the data pyramid explains how the layers trade volume for embodiment closeness.

Stitching combines two hard properties. The work is in the fingers, and the material has no fixed shape. On the first point, the site’s tool use page says that grip changes and tool hand-off between hands live in finger-level detail that RGB video alone cannot resolve, and that each needs 3D hand pose tracked through reach, grasp and tool manipulation. That page is about driving and fastening tools, not needles, so this page does not claim the same findings for stitching. On the second point, the site’s household chores dataset page says deformable objects dominate that data, so instance masks are shipped per frame rather than propagated. Cloth, canvas and thread are the same kind of problem.

The same tool use page says metric depth on sub-centimetre parts is flagged rather than interpolated. Whether a needle or a thread counts as such a part is not stated, and neither is a mask accuracy figure for thread or needles. What carries over is the reasoning: a clip that shows a piece of cloth somewhere in frame does not carry what a policy needs. The hands, the needle and the step boundaries have to be in the episode.

A custom collection lets you aim at the step your model is missing, such as pulling thread through, or following a counted pattern across canvas. This page does not claim that any public corpus lacks stitching data, and it does not claim a training outcome or a model improvement from any dataset. The tradeoff to plan for is the embodiment gap between a human pair of hands and a robot platform doing the same, which retargeting or pairing with robot data has to close.

The site describes the capture setup for the studio and leather rows: a head-mounted phone records the first-person view of real work on site, and a multiview rig adds synchronized third-person angles when the spec asks for them. The lace-threading row on the Deformables & Textiles Bench uses a different setup, a multiview rig with handheld grippers, which the site describes as a staged working set. The site does not explain the grippers further.

The site names no studio, school, workshop, city or region, and it lists capture at your own facility as not offered. This page therefore makes no claim that episodes are recorded in a particular studio, and none that they are recorded at yours. Capture is run to your skill spec using the settings and contributors Firsthand arranges.

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. An undocumented bystander is also one of the machine-checked reject criteria. The anonymization record is delivered with the data.

The "Custom garment fitting" row pins and marks a garment on the client, so a fitting involves a second person. It is not one of the rows on this page. Any stitching spec that would put a second person in frame should settle consent for that person first. On safety, the site states no contributor safety protocol, no injury-prevention guidance and no certification for any stitching row, and this page describes none. 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 say whether the Textile Arts & Stitching Studio environment counts as deep or thin, so ask which timeline applies when you request the quote. Speed comes from a warm pipeline, not from cutting consent, review or anonymization.

The site also states program milestones: a pilot reaches its first delivery in two to three weeks, and a standing program reaches full rate in four to six weeks. After the first batch, on-demand top-ups run against the same accept criteria, so you can add materials, steps or stitch types without rewriting the spec. The site does not state a turnaround for those later batches, so plan them in your quote.

Only footage that clears the signed accept criteria is billed as a validated hour, and rejected items are recaptured at our cost. 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. Collection is fast and on demand in the sense that you do not need a standing program: a focused brief is a normal starting point.

Where this site does not state something, this page says "Not stated" rather than filling the gap. The tables below separate what is delivered from what is not stated. There is no listed label for stitch type, thread type, fabric type, embroidery style or finished-piece quality, no hoop, frame or tool specification beyond the needles and thread named in the catalog rows, no task duration, no site count, no region or city, no episode count and no success or quality label. Nothing here reports a robot result, a benchmark or a model improvement.

Counts of contributors, hours and tasks are set in your skill spec. A task outside the rows quoted above, such as a stitch the catalog does not name, is settled in a Spec Pilot rather than assumed.

Describe the step, the material, 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, a small, fast first batch that proves the spec and the accept criteria on real footage before volume scales. The scoping checklist below maps the parts of a skill spec onto a stitching 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. Decide whether you want hand-only rows such as hand embroidery and needlepoint, or the mixed rows that also involve a machine. And name where the episode starts and stops: one stitch, one line, one motif or the whole piece, since the site does not set an episode length.