Egocentric Wheel and Tire Dressing Data

It is first-person video of a person cleaning the wheels and tires of a vehicle and applying dressing to them, recorded from their own point of view with hand pose, depth and sensor streams alongside. The task is the one the site’s catalog calls wheel and tire dressing: cleaning and dressing wheels and tires with finishing products in one row, and cleaning wheels and tires and applying dressing to the rubber and wheel faces in the other.

The data is human capture by paid, consented contributors, not robot or teleoperation data. The buyer sets the task and the accept criteria in a written spec, and Firsthand handles contributors, capture, QA, consent and delivery.

Two, both rated Easy. "Car Wheel & Tire Dressing" sits in the Car Wash & Detailing environment under the Automotive & Transport sector and is described as cleaning and dressing wheels and tires with finishing products. Its status is Active collection, and the environment lists car wash bays, a detailing shop and polishing stations as examples. "Wheel & Tire Dressing" sits in the Auto Service Center & Tire Shop environment in the same sector and is described as cleaning wheels and tires and applying dressing to the rubber and wheel faces. Its status is Ready to deploy, and that environment lists a car garage, a dealership service bay, oil changes, tire rotation, brake work and vehicle diagnostics as examples.

Both rows list a head-mounted phone or a multiview rig as the capture setup. Active collection means recording is running now, so a spec can join an existing capture stream. Ready to deploy means contributors and site access are set up and capture starts once a spec is signed. The first table below reproduces both rows as the catalog states them. This page names only the surfaces and steps that appear in those rows: wheels, tires, the rubber, the wheel faces, cleaning, dressing and finishing products.

The egocentric tire and wheel service page is about the machine work: dismounting and mounting tires, balancing wheels and rotating them with a pressure check. It lists dressing as one row among them, along with wheel removal inside brake rows, and it has no stereo or MCAP treatment. The egocentric car washing and detailing page covers the whole Car Wash & Detailing vocabulary, from exterior washing to interior vacuuming, and treats the two dressing rows as two of ten.

The stereo car detailing and polishing page is the stereo and MCAP treatment of three exterior rows, and the stereo car interior vacuuming page covers the two interior rows. The stereo tire and wheel service page covers the tire machine rows and does not repeat the dressing rows. This page is the first-person treatment of the two dressing rows alone. It does not repeat the washing, polishing, interior or tire machine rows, and the comparison table below sets the pages side by side.

The catalog names two kinds of surface and two kinds of work. The surfaces are the tire rubber and the wheel face. The work is cleaning and dressing. The site rates both rows Easy, and its tire service page describes dressing as a lighter task than the machine rows, because the hands do surface work rather than heavy handling and the wheel stays in place. A spec should still say which parts an order covers, because cleaning a wheel face, treating the rubber and dressing both are different sequences with different hand positions.

The site states no episode length, no number of wheels per episode, no wheel or tire type and no vehicle type for either row, so those are set per program. Name what one episode is, whether one wheel, one side of a vehicle or a full set, and whether cleaning and dressing sit in the same episode or in separate ones. Then confirm both on a Spec Pilot. The fifth table below shows what each spec element decides.

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. A wrist-mounted camera is available alongside the head camera for close-range detail. That is the capture setup both catalog rows list, and it is what this page describes.

The site does not state a camera count, resolution, frame rate or protection from water and spray for car washing and detailing, so none is claimed here and all are worth raising in the spec. Stereo is a separate offer. The site describes an off-the-shelf stereo head rig, dataset 06, with dual fisheye cameras at 1080p per eye and a 62 mm baseline, shipped as MCAP, but it does not state that the dressing rows are captured with it. A stereo pair for a dressing order would be agreed in the skill spec, and the stereo car detailing pages cover that side of the work.

Each episode follows the Firsthand episode schema: ten streams synced under 2 ms, with head and wrist RGB, depth, 3D hand pose with per-joint visibility, contact and grasp events, and frame-accurate action segments. For dressing work, hand pose tracks the cleaning and the dressing, contact events mark when the fingers meet the wheel, the tire or a tool, and action segments mark each step.

Object labels use the terms your spec sets. Vehicle makes or models, wheel or tire type, product or applicator names, water or pressure settings, device models, throughput, force or tactile data, a finish or gloss measurement and a task-completion label are Not stated. They are defined in the spec if your policy needs them, and no safety or compliance claim is made. The fourth table below sets each stream beside what the site states.

The site states that infrared depth is hostile on painted and oily automotive surfaces, so episodes on those surfaces carry a depth-quality flag instead of being interpolated. A buyer can then keep, weight or drop the flagged frames rather than train on filled-in values that look real.

The site does not state how depth behaves on a wheel face, on tire rubber, or on a wet or glossy finish, and it gives no depth accuracy figure for any automotive row. This page therefore promises no depth figure and makes no claim about how a finished wheel looks in depth. If depth matters to your policy, say so in the spec and check it on the Spec Pilot batch.

Hands are the first risk. The hands guide a tool or a cloth against a round surface, so the contact point is often hidden by the wheel itself. The visibility flag on each joint is the answer the site gives, and contact events mark when a hand or tool meets the part. A spec can also ask for wrist views, so a grip the head camera loses is still recorded, and for a multiview rig when a second angle is wanted.

Step order is the second. Cleaning and dressing look similar from a distance, and action segments are what separate them, so the spec should say which boundary marks the end of one and the start of the next. Surface depth is the third, covered above by the depth-quality flag.

Capture quality is the fourth. The site states that an episode is rejected when the manipulated object is out of frame for more than 15 percent of it or an undocumented bystander appears, and that a reject log ships with every batch. Rejected items are recaptured at our cost.

Delivery formats are RLDS, LeRobot and WebDataset, plus HDF5, zarr and Rerun .rrd. Firsthand also offers MCAP delivery, and the MCAP hub page covers what the site states and does not state about it. Dressing episodes use the same schema as every other Firsthand collection, so a loader that reads another Firsthand dataset reads this one without changes.

Every batch is validated against signed accept criteria before it ships, with a reject log, and delivery is rolling, per validated batch. The sources section lists the official documentation for the three training formats named here.

A car wash bay or a service bay is a shared space, so bystanders matter, and a vehicle can carry a plate. For every delivered episode the site states that faces, plates and screens are detected and irreversibly blurred at ingest, that a reviewer confirms the result before anything ships, and that footage that cannot be de-identified is discarded. The anonymization record is delivered with the data, and every episode links to a signed contributor release through a documented consent chain.

The site does not state a separate rule for other vehicle text such as badges or stickers, so this page does not claim it is blurred. Raise it in the spec if it matters. The site also states no contributor safety protocol and no guidance on cleaning or finishing products for either row, so this page describes none, and it does not state capture at a customer’s own facility.

Where coverage is already deep, the median time from signed spec to first validated batch is a median of 48 hours; where it is thin, turnaround runs three to four weeks, because vetting operators there is the bottleneck, and we quote that up front. The site does not say which applies to the dressing rows, so ask when you request the quote. The car wash dressing row is Active collection, which is the faster route, and the auto service row is Ready to deploy.

This is custom collection: custom automotive detailing data collection at scale through a network of thousands of capture providers, for a consistent, recurring data flow to labs. On-demand top-ups run against the same spec and accept criteria after each training run. Collections are billed per validated hour, and no delivered hours are claimed for either row.

Describe the wheels and tires, the setting, the steps in order, what counts as one completed wheel, and the variation your pipeline is missing. Then name the streams you want, the delivery format your loader expects and whether a second angle or a wrist view is needed.

Most programs begin with a Spec Pilot, a small first batch that proves the spec and accept criteria on real footage before you scale. The fifth table below shows what each spec element decides.