Egocentric Trash and Recycling Task Video Data

Egocentric trash and recycling task video data is real-world, first-person video of a person handling waste as part of a working day: pulling full bags and recycling out of a kitchen, relining the cans, bagging trash at a counter, sweeping a venue and bagging what is collected, and carrying the bags to the bins or dumpsters. It is recorded from the operator viewpoint, with depth, 3D hand pose, contact events and frame-accurate action segments aligned on one hardware clock.

These tasks look simple and still ask a lot of a manipulation policy. A bag has no fixed shape, it has to be gripped by whatever part of it is reachable, and it is carried from one place to another before it is dropped into a bin. A first-person view keeps the bag, both hands and the destination in the same frame, which is the viewpoint a robot arm with a head or wrist camera would act from.

This page treats trash and recycling as a task family in its own right. Sorting is about moving items to destinations and cleaning is about wiping and mopping surfaces. Here the handled object is waste: collected, bagged, carried and disposed of.

The task catalog already lists trash and recycling work in several settings. In a restaurant and commercial kitchen it includes pulling full bags and recycling from the kitchen, carrying them to the bins and relining the cans. At an event venue it includes sweeping the venue floor, bagging trash and recycling and carrying the bags to the dumpsters, and a second row that walks the venue collecting trash and recycling into bags and sweeping up debris. In a working kitchen set it includes clearing and wiping a counter and bagging trash, and in a hotel lobby it includes emptying bins as part of common-area cleaning.

The catalog also lists nearby rows that touch waste without being trash handling as their main task. A listed recycling and waste-transfer facility environment names sorting lines, baling and material recovery. E-waste disassembly sorts materials into bins, damaged-goods sorting ends with an item placed in a markdown or disposal bin, and construction cleanup includes clearing debris and sorting waste.

A custom spec is scoped from that same vocabulary. You describe the bags, bins, stations and sequence your policy needs, and the written spec covers them, with the catalog tasks as a starting point rather than a fixed menu. The first two tables list every row with its description, environment and status as the catalog states them.

The clearest trash row in the catalog is trash and recycling removal in a kitchen. It is described as pulling full bags and recycling from the kitchen, carrying them to the bins and relining the cans. The row is rated Easy, is recorded with a head-mounted phone and a multiview rig, and is Ready to deploy.

It sits in the Restaurant & Commercial Kitchen environment, whose examples are a sit-down restaurant kitchen, a cafe, a coffee shop, an artisan bakery and a patisserie. The site does not state which specific kitchens a collection runs in, so the environment, access and schedule are agreed in the written spec.

A second kitchen row is staged rather than recorded on site. Wipe counter and trash clears and wipes a counter and bags trash, in a Working Kitchen Set described as a stovetop, counter, sink, dishwasher, utensils and food props. It is recorded by a multiview rig while the operator uses handheld grippers, and it is Available on request.

Two Event Venue rows are the most direct statement of recycling alongside trash. The first is described as sweeping the venue floor, bagging trash and recycling and carrying the bags to the dumpsters, and it names brooms and liner bags as the tools. The second, Trash & Recycling Sweep, is described as walking the venue collecting trash and recycling into bags and sweeping up debris.

Both are rated Easy, both are recorded with a head-mounted phone and a multiview rig, and both are Ready to deploy. The Event Venue environment lists conference halls, wedding venues, stages and banquet rooms as examples.

These rows combine moving through a space while collecting with a carry to a disposal point. A spec can ask for either on its own or the full sequence, and action segments mark where collecting stops and carrying starts.

Bin emptying appears inside a larger cleaning row rather than as a task of its own. Lobby and common-area cleaning in a hotel is described as vacuuming and mopping lobby floors, dusting furniture, polishing glass and emptying bins. It is rated Easy, recorded with a head-mounted phone and a multiview rig, and Ready to deploy.

Because emptying bins is one step among several in that row, a spec that needs bin emptying as the main task says so, and the action segments isolate that step from the vacuuming, mopping and dusting around it. The site does not state how many bins or how much waste an episode covers.

Counter trash works the same way. In the working kitchen set, bagging trash is the last step of a clear and wipe sequence, so the wipe and the bagging are separate action segments.

Recycling appears in the catalog as an environment: Recycling & Waste-Transfer Facility, with sorting lines, baling and material recovery as examples. It is described only as a target environment for commercial egocentric collection, and it is rated Easy, recorded with a head-mounted phone and a multiview rig, and Ready to deploy.

The catalog lists no separate task row under that environment. So the specific recycling sorting tasks, such as which materials go to which stream, are scoped in the spec rather than stated as catalog tasks, and this page does not promise a recycling sorting video dataset with a fixed material list. The site does not state recycling streams, material categories or municipal rules.

Two nearby rows sort waste. E-waste disassembly and sorting, which is rated Medium and Available on request, unscrews and pries open devices, removes boards and batteries and sorts materials into bins. Damaged-goods sorting in a retail back of house inspects returned or damaged items, scans them and sorts them into markdown or disposal bins, and it is Ready to deploy. The general sorting task is covered on its own page, linked below.

The catalog places these rows in a restaurant and commercial kitchen, an event venue, a staged working kitchen set and a hotel and lodging environment. The adjacent sorting and debris rows add a recycling and waste-transfer facility, an electronics manufacturing and e-waste environment, a retail back of house and an active construction site. Each task row carries a status that says how soon capture can begin.

The site defines those statuses plainly. Ready to deploy means contributors and site access are set up and capture starts once a spec is signed. Active collection means recording is running now, so a spec can join an existing capture stream. Available on request means the work is scoped and recruited to spec when a buyer asks for it. In the current catalog the kitchen removal row, both venue rows and the hotel lobby row are Ready to deploy, the working kitchen set row is Available on request and the construction rows are Active collection.

The site does not state a named waste-management company, municipality or recycling program for any row, and it does not state capture at a customer own facility. Environment, access and schedule are agreed in the written spec.

The catalog names two capture modes. In the first, 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. In the second, a staged working set is recorded by a multiview rig while the operator performs the task with handheld grippers, so demonstrations map onto a robot end effector.

A head-mounted camera keeps the pull, carry and drop motion in frame from the viewpoint a manipulation policy will see from, with a wrist-mounted camera available alongside it for close-range detail. Depth, 3D hand pose, contact events and action segments are aligned to video on one hardware clock, synced under 2 ms. Distance-invalid pixels beyond the depth sensor reliable range are flagged rather than interpolated.

The site does not state a camera count, resolution or frame rate specific to trash handling, and it does not state a bin or bag size, so those are agreed in the written spec rather than assumed.

Each episode follows the same episode schema used across Firsthand custom collections: 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 trash and recycling, hand pose tracks the operator through the grip, lift, carry and drop, contact events mark when the fingers make and break contact with a bag, a bin or a liner, and action segments mark the start and end of each step in the sequence. Because the labels share one clock with the video and depth, a step boundary lines up with the exact frame where it happens.

Object labels use the terms your spec sets. The skill spec decides which object classes are in scope before capture, so labels follow the vocabulary your training stack expects rather than a generic taxonomy.

Bin or bag sizes, waste volumes or throughput, a material-classification label or accuracy, recycling-stream or municipal rules, force or tactile data, a task-completion label and any hygiene or safety certification are not stated anywhere on the site as standard fields. Some catalog task names mention liner bags, bins or dumpsters, but those describe what the operator does, not a field delivered with the episode.

If your policy needs any of these, they are defined in the written spec as annotation depth, and the quote covers them before capture begins. The page does not promise them by default, and it makes no hygiene, safety or environmental claim.

Delivery formats are RLDS, LeRobot and WebDataset, plus HDF5, zarr and Rerun .rrd. The first three are the training-loader formats; HDF5 and zarr carry arrays for custom pipelines, and Rerun .rrd gives a multi-stream synchronized playback file for review. The format table below sets each against what its official documentation states.

Converter source is readable, so you can retarget the schema to your own loader, and the episode schema page lists the fields. Trash and recycling episodes use the same schema as every other Firsthand collection, so a loader that already 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.

Every episode links to a signed contributor release through a documented consent chain, and the consent record is part of the delivery. Contributors are the people who perform the trash and recycling tasks, so their release covers the footage captured of their hands and workspace.

Faces, plates and screens are detected and irreversibly blurred at ingest, and a reviewer confirms the result before anything ships. The anonymization record is delivered with the data. A venue, a hotel lobby or a kitchen can put other people in frame, which is why that pass matters here.

Text on bags, bins or discarded items is not listed as a blur category on the site, so this page does not claim it is blurred. If text on waste matters for your use, raise it in the spec so the handling is agreed before capture.

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 timeline up front. The catalog status on each task row tells you which case you are in.

On-demand top-ups run against the same spec and accept criteria after each training run, so new trash and recycling data stays consistent with what you already hold. The site does not state a batch size or a later-batch turnaround, so those are quoted against your spec.

You describe the bags, bins, stations 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 against that document before any capture begins.

A Spec Pilot, a small and fast first batch, can prove the spec and the accept criteria before you scale volume. The collection runs through the same seven-stage end-to-end custom collection pipeline as every Firsthand program: contributor recruiting, capture, per-batch QA, anonymization and licensed delivery. Billing is per validated hour, a reject log ships with every batch, and rejected items are recaptured at our cost.

Every custom collection ships under a perpetual, buyer-owned commercial license, and exclusivity is available so the data is never resold. The license is documented alongside the consent chain, so a policy trained on the data can go into a product.