Stereo Security Checkpoint Screening Data

It is real-world, first-person video of the hand work done around a staged checkpoint, recorded with two calibrated cameras so that depth can be recovered, together with 3D hand pose and the other streams the site already lists. The catalog describes the work in short phrases: place bags and loose items into bins, load bags to a belt at pitch, open, inspect and re-pack a bag, retrieve a flagged item from a bin, and stack empty bins and return them. The buyer sets the task and the accept criteria; Firsthand handles contributors, capture, QA, consent and delivery.

The data is human capture by paid, consented contributors, not robot or teleoperation data. This page is about security checkpoint robot training data in the sense of what a policy can learn from watching a person handle trays, bins and bags, and it makes no claim that any robot has been trained on it or that any robot screens anything. People reach this topic through several phrasings, such as stereo baggage screening video, an egocentric screening tray and bin handling dataset or custom security-operations data collection at scale. They describe one request from different angles.

No. The catalog environment is named Mock Security Checkpoint, and its examples line reads "Screening bins, a belt, a wand, a mannequin, bag props (no live airport)". This is a staged working set, and this page repeats the site wording instead of adding to it. It does not describe capture at a real airport, agency or operator, and it does not describe real travellers or passengers being recorded.

Where a row involves a person-shaped target, the catalog names a mannequin. The site names no airport, agency, operator or client for this environment, and it states no screening equipment model, detection method, threshold or prohibited-item list. Those are Not stated, and the page claims no screening technology and no regulatory or agency standing.

Twelve rows sit under Mock Security Checkpoint in the Public Spaces & Outdoor sector. Nine are rated Easy and three are rated Medium, and every row has the status Available on request. That means the work is scoped and recruited to spec when a buyer asks for it, so this page claims no completed collection, no delivered hours and no delivered stereo footage for any of them. The first table below reproduces all twelve rows exactly as the catalog lists them.

The rows fall into four groups. Trays and bins: Load screening trays, Retrieve flagged item and Bin stacking + return. Bags and packages: Baggage load to belt, Bag search + re-pack, Package inspection and Divest + re-collect. Handheld and mannequin: Handheld wand pass and Pat-down sequence (mannequin). Badges, tags and documents: Access badge + gate, Baggage tag scan and Document check. The stereo pages in this cluster care most about the first two groups, because those rows are hand-and-object work with containers that open and shift.

Load screening trays is Easy and reads "Place bags and loose items into bins." Retrieve flagged item is Easy and reads "Retrieve a flagged item from a bin." Bin stacking + return is Easy and reads "Stack empty bins and return them." Together they form a loop of filling, emptying and resetting. The site does not say what makes an item flagged, and it states no bin sizes, item lists or fill rules, so a spec should name the item groups, how many go in a bin and which item is retrieved.

Baggage load to belt is Medium and reads "Load bags to a belt at pitch." Bag search + re-pack is Medium and reads "Open, inspect, and re-pack a bag." Package inspection is Medium and reads "Open, inspect, and reseal a package." Divest + re-collect is Easy and reads "Divest belongings and re-collect after." The phrase "at pitch" is catalog wording, and the site states no belt speed, spacing or bag counts. What inspecting means is a spec decision. This page gives no procedural detail beyond these row names.

Handheld wand pass (Easy) reads "Pass a wand over a mannequin." Pat-down sequence (mannequin) (Easy) reads "Perform a pat-down pattern on a mannequin." Both use a mannequin and no real person, and the site states nothing more about them, so this page says nothing more either.

Access badge + gate (Easy) reads "Handle a badge and gate object." Baggage tag scan (Easy) reads "Scan a bag tag and route." Document check (Easy) reads "Place a document under a reader." The catalog gives one line for each. The site states no reader, tag or badge types, and it states no document formats or routing rules, so those are agreed in the skill spec if an order includes these rows.

Not stated. The catalog lists one capture setup for all twelve rows, a multiview rig with handheld grippers, which records the staged working set while the operator works. It does not list a stereo pair. Stereo applies here only through the site general stereo offer: the site describes a stereo head rig and a hard-environments stack, and a stereo pair for a checkpoint order is a choice made in the skill spec and confirmed on a Spec Pilot.

This is why the page is written as a stereo and container page for these rows, not as a description of existing stereo checkpoint footage. If you need stereo, say so in the brief. If you are happy with the catalog setup, the egocentric security checkpoint page describes that non-stereo route for the same twelve rows.

This page is stereo capture of a staged checkpoint working set: trays, bins, bags, tags and belt loading. The egocentric security checkpoint page covers the same twelve rows as first-person video, with stereo not its subject. The egocentric luggage and baggage handling page covers four rows across a hotel and the mock checkpoint, and it is not stereo.

The stereo warehouse sorting and labeling page and the stereo order packing and taping page are stereo, but they cover warehouse rows such as label printing, box sorting and order packing, in a different environment and at a different status. Neither covers trays, screening bins or the mock checkpoint. The second table below puts the pages side by side with their status as the site lists it.

The catalog states one mode for all twelve rows: 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. The site describes two stereo configurations. The stereo head rig, dataset 06, is a dual fisheye pair at 1080p per eye on a 62 mm baseline, with per-unit calibration, a 9-axis IMU and hand-tracking output, shipped as MCAP. The hard-environments stack is a 4K/60 head camera with a 120° lens, two 1080p/60 wrist cameras, depth at 848×480 and 30 fps and a 200 Hz 6-DoF IMU. Which of these a checkpoint order uses is Not stated and is agreed in the skill spec.

Whichever rig is chosen, calibration is measured every session and shipped with the episode, with a 0.28 px RMS reprojection gate. Hand pose is 21 joints per hand with a per-joint visibility flag, and the language data is one task instruction per episode. A head-mounted first-person view is not listed for these rows, so whether one is added is settled in the spec. Camera count, resolution and frame rate for the set as quoted are Not stated.

Hand occlusion is the first case. Fingers close around an item as it goes into a tray, and a bag held open hides what is inside it. The site states that the hands occlude the work and that fingers disappear behind objects. The per-joint visibility flag marks joints that are not visible instead of guessing them, contact events mark when a hand meets a bin edge, a bag or an item, and a spec can ask for wrist or multiview angles so a grip the head camera loses is still recorded.

Reflective surfaces are the second. The site states that depth degrades on reflective surfaces, which per-frame confidence maps handle by masking low-confidence regions. How depth behaves on a transparent bin or on a particular bag prop is Not stated, so a Spec Pilot checks it on real footage. Deformable bags are the third: the site states that masks for deformable objects are shipped per frame rather than propagated, with instance masks stored at 848×480, and what counts as the object, the bag or its contents, is a spec decision.

Belt motion, crowded tables and lighting are the rest. All streams share one hardware clock with a 2.0 ms ceiling, so a bag can be followed onto a moving belt across every stream, but the site states no belt speed. It says nothing about table crowding, and an undocumented bystander is a stated reject criterion. The site states no lighting values for this set, so conditions are fixed in the skill spec.

Every batch is validated against signed accept criteria before it ships, with a reject log. The site states that an episode is rejected when the manipulated object is out of frame for more than 15 percent of it, when more than 35 percent of depth pixels in the manipulation region are invalid, when a sync breach runs beyond the ceiling for more than three consecutive frames, when an undocumented bystander appears, or when the episode does not match the named skill spec. Episodes that fail are rejected and logged per batch, not delivered with a caveat.

Billing is per validated hour, and rejected items are recaptured at our cost. The site does not state an accuracy figure for step labels, and it lists no label for item identity, bag contents, success or quality. What counts as a pass for a step boundary, such as the end of a tray load or the start of a retrieval, is set in the accept criteria.

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, and a reviewer confirms the result before anything ships, with the anonymization record delivered alongside. Footage that cannot be de-identified is discarded, and an undocumented bystander is a stated reject criterion. The mannequin rows keep the person-shaped target free of any real person.

The site states no checkpoint-specific rule for how documents under a reader, badges or tags are blurred, so that is Not stated beyond the general rule above, and it is raised in the spec if it matters. The buyer owns the dataset under a perpetual, irrevocable, non-exclusive-by-default licence, with exclusivity available, and the executed agreement controls. The privacy and anonymization pages are linked below.

Firsthand offers MCAP delivery. MCAP is an open-source container format with its own official site at mcap.dev, a specification at mcap.dev/spec and a repository at github.com/foxglove/mcap; read those for what the container is and how it stores records, timestamps and attachments. This page describes the container only from those sources and does not state anything more about it.

The streams the site already states for a stereo capture are the stereo pair, depth with confidence maps, 21-joint hand pose, IMU, contact events, action segments, per-frame instance masks and calibration, with one task instruction per episode. Topic names, message schemas, encoding and rates are Not stated on the site and are agreed in the skill spec. Only dataset 06 is stated as shipped as MCAP; for any custom order, MCAP can be the only delivery or sit beside RLDS, LeRobot, WebDataset, HDF5 or zarr. The third table below maps each stream to what the site states.

Airport, agency, operator, client and brand names, screening equipment models, detection methods or thresholds, prohibited-item lists, bag sizes and contents, item counts, belt speed, task times, throughput, error rates, force or tactile data and any regulatory or certification standing are not stated on the site. None is claimed here.

If your policy needs any of the fields that can be specified, they are written into the skill spec as annotation depth and covered in the quote before capture begins.

The site off-the-shelf catalog lists about 324,800 hours across 18 datasets, including about 64,800 hours of stereo video in 9 datasets. None of its dataset descriptions names checkpoint, tray, bin or baggage tasks, so no off-the-shelf set stands in for these rows. One stereo set is published as a format and rig example, and it is not coverage of any checkpoint row.

Stereo checkpoint collection to a spec is therefore custom collection: custom collection at scale through a network of thousands of capture providers, for a consistent, recurring data flow to labs. Where coverage is 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. All of these rows are Available on request, so the scoped-to-spec timeline applies, and the quote states it up front.

Describe the rows you want, the steps you expect to see and where each starts and ends. Say which item groups go in a tray, whether an episode is one tray, one bag or a full loop of load, retrieve and return, what counts as the object, what counts as a reject, and whether you want a stereo pair, wrist views or the catalog multiview rig. Then name the streams you want in the file, the topic naming and message encoding your reader expects, whether calibration ships inside the file or beside it, and whether MCAP is the only delivery.

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 last table below shows what each spec element decides.