Egocentric Food and Beverage Manufacturing Video Data

Egocentric food and beverage manufacturing video data is real-world, first-person video of work in a plant that makes packaged food or drink. It is recorded from the worker viewpoint, with depth, 3D hand pose, contact events and frame-accurate action segments aligned on one hardware clock. The tasks it covers are the ones the site states in its task catalog, under the Food & Beverage Manufacturing Plant environment: inspecting and removing defective products, operating and monitoring a filling line, packaging, sealing and labelling, running a production line, foreign-object inspection and sanitation washdown.

The produce packing page covers fruit sorting and packing on a farm, and the print and packaging page covers print and packaging work. Neither is about a manufacturing plant that fills, seals and labels food and drink. This page covers the plant sequence, from a product arriving on the line to a sealed, labelled pack and a washed-down production area.

The site states plant work in one catalog environment, Food & Beverage Manufacturing Plant, under the Food & Beverage sector. Its examples line reads "Packaged food factory, bottling plant, meat processing facility, ready-meal production line." Eight task rows sit under it, four rated Easy and four rated Medium, and every one lists a head-mounted phone or multiview rig as the capture mode.

All eight rows carry the status Ready to deploy. The site defines that status as contributors and site access set up, with capture starting once a spec is signed. This page uses all eight rows and lists them in the table below as the catalog states them.

Two rows are about looking at product and removing what should not go forward. The first reads "Inspect and remove defective products using an inspection light and gloves," described as holding products under the inspection light, spotting defects and pulling them from the line. The second is QA / Foreign-Object Inspection, described as inspecting items on the belt for foreign objects and defects and removing anything suspect.

Both are rated Easy. What counts as a defect or foreign object, and the product inspected, are Not stated and are set in the written spec. The gloves in the first row are part of the task and do not imply tactile data.

Operate and monitor a filling line using control panels and gauges is described as starting the filling line, reading gauges, adjusting settings on the panel and clearing minor jams. Production Line Operation is described as feeding product onto the line, monitoring the machines and packing finished goods at the end. Both are rated Medium.

These are the rows where the hands work a control panel as well as the product. Machine names, line speeds and panel layouts are Not stated.

Packaging & Sealing is described as filling packages, running them through the sealer and applying labels before casing. A second row, "Seal and label packaged goods using a heat sealer and label applicator," is described as feeding packs into the heat sealer, checking the seal and running them through the label applicator. Both are rated Easy.

Between them they cover the end of the line: fill, seal, check the seal, label and case. The site does not state pack formats, pack counts or label contents. It also does not describe a seal check as a recorded quality label, so no seal-quality outcome is claimed here.

Sanitation Washdown is described as rinsing equipment with hoses, applying foam sanitizer, scrubbing and rinsing down the production area. The longer row, "Wash down and sanitize production equipment using hoses and foam sanitizer," is described as disassembling guards, hosing down equipment, foaming, scrubbing and rinsing to a sanitary finish. Both are rated Medium.

These are the wet, multi-step rows, and they are included because the catalog states them. "Sanitary finish" is catalog wording and not a standard: the site states no food-safety, hygiene or cleaning certification, and none is claimed here. Any acceptance rule for a washdown is written into the spec.

The catalog places the work in a food or beverage plant, and the environment page states that contributors are recruited to spec in Western program markets, so the footage reflects how this work is done in North American and European businesses. The site names no plant, company, city or region, and it does not state capture at a customer own plant. Access and schedule are agreed in the spec, and the number of plants is Not stated.

The privacy steps the site states are these. Every episode links to a signed contributor release through a documented consent chain, and the consent record is part of the delivery. Faces, plates and screens are detected and irreversibly blurred at ingest, a reviewer confirms the result before anything ships, and the anonymization record is delivered with the data. Physical-safety measures, PPE rules and plant access terms are Not stated.

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. All eight plant rows list this single setup. On the rig tier, every stream is aligned to a shared clock and checked at the sync gate before an hour is billed.

Camera count, resolution and frame rate for plant work are Not stated, so session length and volume are quoted against your spec.

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. In a plant, hand pose tracks how the fingers hold and place each item and work a panel, contact events mark when fingers meet a product, a pack or a control, and action segments mark each step.

Object labels use the terms your spec sets. Force or tactile data, and defect, seal or quality-outcome labels, are Not stated for these rows. If your policy needs them, they are defined in the spec. Tactile gloves are a separate capture tier with their own page.

Delivery formats are RLDS, LeRobot and WebDataset, plus HDF5, zarr and Rerun .rrd. The format table below sets each against what its official documentation states. Action segments are the step labels, and any further object labels follow the spec.

Every batch is validated against signed accept criteria before it ships, with a reject log, and delivery is rolling, per validated batch. Billing is per validated hour, and rejected items are recaptured at our cost.

A survey of more than 150 studies states that, compared with rigid objects, deformable object manipulation poses significant challenges for perception, modeling and manipulation, because deformable objects have an infinite-dimensional state space and complex dynamics. The same survey states that data-driven approaches can address some of those challenges. That is the case for demonstrations recorded from the worker viewpoint, where the hands and the product are in view together.

The site does not state which products are deformable, so this page makes no claim about any product. It does state wet washdown rows and belt-side inspection rows. Wet work and item-after-item handling are hard to cover with a few recordings, which is why a spec names the tasks, the products and the number of sequences. Cycle times and repetition rates are Not stated.

Where coverage is already deep, the first validated batch lands in a median of 48 hours from signed spec. Thin environments typically take three to four weeks, because vetting operators there is the bottleneck, and we quote that up front. The site does not state which of the two applies to plant work, so the timeline is quoted against your spec.

On-demand top-ups run against the same spec and accept criteria after each training run.

You describe the plant tasks, products and sequence your policy needs. We turn that into a written spec covering tasks, environments, capture tier, annotation depth and accept criteria, then quote reach, timeline and price before any capture begins.

A Spec Pilot, a small and fast first batch, can prove the spec and accept criteria before you scale. Billing is per validated hour, a reject log ships with every batch, rejected items are recaptured at our cost, and every collection ships under a perpetual, buyer-owned commercial license.