Egocentric Dental Clinic Task Video Data

Egocentric dental clinic video data is first-person video of a person doing clinic tasks, 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 the chair and tray, so the frames show what the hands are doing rather than a distant view of a room. Teams searching for a dental operatory video dataset or dental clinic robot training data are usually looking for this: footage of the hands doing the work, tied to labels they can train on.

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 covers the Dental Clinic environment. It is different from the dental lab page, whose bench work is fabricating prostheses and handling instrument sets, and from the broader healthcare page.

The task catalog lists a Dental Clinic environment under Healthcare & Life Sciences. Its examples are a dental clinic, an orthodontic practice and an oral surgery center. It has five task rows. Instrument handling passes, handles and returns dental instruments to the clinician during a procedure, and it is rated Medium. Patient preparation seats the patient, fits the bib and eyewear, adjusts the chair and light and lays out supplies, and it is rated Easy.

Post-procedure cleaning clears the tray, disposes of sharps and waste and wipes and disinfects the chair and surfaces, and it is rated Easy. Suction and irrigation assistance holds the suction and irrigation tools to keep the field clear during treatment, and it is rated Medium. Tray setup lays out sterile instruments and materials on the tray in procedure order, and it is rated Easy. Instrument tray setup data is therefore a named row, not an extrapolation.

All five rows are recorded with a head-mounted phone and a multiview rig, and all five are Available on request. That means the work is scoped and recruited to spec when a buyer asks for it, rather than running as an open capture stream today. The catalog does not split the rows into sub-steps, and it names no procedure types, instrument lists, room layouts 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 a head-mounted phone with a multiview rig for these five rows, while the ten-stream stack is the instrumented tier. Which applies to a dental clinic program is not stated and 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 clinic work, a segment might be one tray layout, one pass of an item to the clinician or one wipe of the chair, as written in the skill spec.

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.

Several of the catalog tasks depend on order and on other people. A tray is laid out in procedure order, so a model has to learn sequence as well as placement. Instrument handling is a pass to a clinician, so the hand-to-hand moment sits at the edge of the frame. Suction and irrigation assistance keeps tools close to the work for long stretches, which is why hand pose carries a per-joint 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, and says so where nothing is stated.

A robot foundation model paper on arXiv organizes its training data as a pyramid, with large quantities of web data and human videos at the base, synthetic data in the middle and real robot data at the top. The same paper says human videos have no explicit action labels, so it learns latent actions to label them. That is robot-model research, 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. Whether a given program uses a working clinic or a staged set is not stated, and neither is the venue type. Both are agreed in the skill spec before capture begins.

Two task rows mention a patient: seating the patient, and passing instruments during a procedure. The site does not say whether a live patient, a volunteer or a stand-in is involved, and this page does not describe capturing patients. That is agreed in the spec, and an undocumented bystander is a reject criterion.

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.

The catalog notes HIPAA compliance for its Hospital & Clinical Care environment, but states nothing like it for the dental clinic. This page therefore makes no clinical, sterilization, infection-control or regulatory claim. Blurring of charts, labels or screens beyond the listed categories is not stated.

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. All five dental clinic rows are Available on request, so the range that applies is confirmed when the spec is quoted rather than assumed.

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. These are program milestones, separate from the per-batch figures above.

Describe the clinic tasks and their order, what counts as a finished step, who may be in frame, and the 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 procedure type, instrument list, room layout, venue type, region, cycle time, number of sites or number of episodes is stated for dental clinic work. There is no quality or success label, no force or tactile stream, no clinical or regulatory status and no robot result. The delivered-versus-not-stated table below separates what is stated from what is not. Nothing here claims a training outcome.