Why does identity come before age?

Because a video cuts. Box tracking followed a cut between an interviewer and a child and reported one person whose age was the median of three people. Every individual estimate was right and the identity was wrong, which is worse because it looks correct. The template groups by face across cuts first, then ages the person, and reports how many looks the estimate rests on.

What does zero missed minors cost?

On single-frame evidence, a cutoff of 18 catches 82% of minors and sends 4.8% of adults to review. A cutoff of 22 catches 94% and sends 23%. A cutoff of 30 catches 100% and sends 94.5% of adult footage to review. The number is yours to choose, and the page shows the curve so you choose it with the price attached.

Does face size limit this?

Less than expected. Shrinking labelled faces from 200px to 32px moved recall from 82.2% to 86.2%. Detection is the real limit: a face the detector never fires on is never aged, which is why anything undetected is routed to review rather than assumed adult.

How do I know the feedback loop is working and not just collecting?

Because the loop ends in an evaluation, not in a collection. Annotations and interactions are turned into a held-out set and a threshold change is measured as precision and recall moving on it. Captured signals that do not move a number are not a working loop.

Can I use your reference sets instead of supplying my own?

You can start from the marketplace Celebrity Face Scanner and Logo detection indexes to see the shape of the results. For a production decision the reference set should be yours, because it decides which marks and which people count, and it is the thing that decays if nobody owns it.

Safety and compliance

Template v1.0 · updated 2026-09-03

Video Moderation for Training Sets

Reduce raw footage to the subset you can defend training on. Every excluded clip carries its reason, every uncertain one goes to a person, and the thresholds are yours.

A cleared training set, with an audit trail for every clip that did not make it.

Deploy this template

Teams preparing video corpora for model training who have to prove what was excluded and why.

82%

of minors caught at a cutoff of 18

on 650 labelled faces, 320 minors and 330 adults, single frame; 4.8% of adults sent to review

100%

recall is reachable, and the page says what it costs

a cutoff of 30 catches every minor and sends 94.5% of adult footage to review; the whole curve is in the evaluation tab

0.70

face detector gate

cleared six phantom minors the detector found in non-faces while keeping 85% of genuine faces

39%

of detections routed to review as unaged

no usable face, under 24px, or detector confidence under 0.70. None of them is passed as adult.

What it looks like

Frame in, decisions fire, a verdict lands, a reviewer's call feeds back.

Simulated walkthrough · illustrative frames, scripted decisions

A close-up interview frame with one face in shot

faces -> age-estimation -> person-verdictsStill: Pexels / Ron Lachframe 1 of 4

Decision path

face detected 0.93
grouped across 3 looks
age estimate: borderline
route: review

within 6 years of the cutoff, and 3 looks

Running tally

cleared

review

excluded

Reviewer feedback

When the frames are done, a reviewer's call on the borderline case feeds back into the thresholds.

What goes in

source-video
Your footage, read from your own object storage. Nothing is copied out.
brand-mark-refs
Logo and mark variants you supply, one file per variant, with a hard-fail or contextual rule each.
person-refs
Public-figure reference imagery you supply and hold the rights to. Several poses per person.

Where the decisions fire

Illustrative frames; boxes show the decision path, not live model output

brand mark · match 0.88 · rule: contextualcleared

keyframes -> brand-mark-match · A keyframe embedding is matched against your brand-mark reference index. The rule attached to that mark decides hard fail or contextual.Still: Pexels / Mathias Reding

Two people sparring in a gym — keyframes -> violence-context · Violence is scored in context. Sparring, practice and play score staged; the real thing scores real. The margin between them is what you set.Still: Pexels / Franco Monsalvo

A reporter beside a van carrying a BREAKING NEWS graphic — onscreen-text · Words shown on screen and never spoken are read by the on-screen text pass, static overlays and scrolling bands alike, and land in their own searchable field. Nothing in the audio says BREAKING NEWS.Still: Pexels / cottonbro studio

What comes out

segment-verdicts
One row per segment: decision, the reason, and counts of people, minors, unaged and borderline faces.
review-queue
Everything routed to a human, ranked, with the decision chain attached to each item.
cleared-set-export
The segments you can train on, as a filterable set ready to hand to a training job.

How the namespace is wired

3 buckets, 8 collections, 2 clean views, 6 retrievers. The diagram generates the manifest below; they cannot drift apart.

SourceBucketCollectionClean viewRetrieverClick a node to inspect it

DocsNamespaces Pipeline configuration Retrievers

One file spins up the namespace. Generated from the diagram above. Also served at /templates/video-moderation.namespace.yaml.

# One file spins up the namespace. Wiring comes from the diagram:
# edges are sync targets, source.bucket_ids, view sources and retriever scope.
namespace: video-moderation
sources:
  - name: source-video
    kind: s3
    cadence: continuous
    filters: your footage, in your account
    syncs_into: [source-video]
  - name: brand-marks
    kind: manual
    cadence: one-shot
    filters: customer-supplied logo and mark variants
    syncs_into: [brand-mark-refs]
  - name: person-references
    kind: manual
    cadence: one-shot
    filters: customer-supplied public-figure reference imagery
    syncs_into: [person-refs]
buckets:
  - name: source-video
    skip_duplicates: true
    schema:
      content: { type: video }
  - name: brand-mark-refs
    skip_duplicates: true
    schema:
      content: { type: image }
  - name: person-refs
    skip_duplicates: true
    schema:
      content: { type: image }
collections:
  - name: video-segments
    source: { type: bucket, bucket_ids: [source-video] }
    feature_extractor:
      feature_extractor_name: multimodal_extractor
      version: v2
      input_mappings:
        video: content
      field_passthrough:
        - { source_path: file_location }
        - { source_path: segment_id }
        - { source_path: start_time }
        - { source_path: end_time }
  - name: brand-mark-index
    source: { type: bucket, bucket_ids: [brand-mark-refs] }
    feature_extractor:
      feature_extractor_name: image_extractor
      version: v1
      input_mappings:
        image: content
      field_passthrough:
        - { source_path: mark_name }
        - { source_path: rule }
  - name: person-ref-index
    source: { type: bucket, bucket_ids: [person-refs] }
    feature_extractor:
      feature_extractor_name: face_identity_extractor
      version: v1
      input_mappings:
        image: content
      field_passthrough:
        - { source_path: person_name }
        - { source_path: rights_note }
  - name: faces
    source: { type: collection, collection_id: video-segments }
    feature_extractor:
      feature_extractor_name: face_identity_extractor
      version: v1
      input_mappings:
        video: content
      field_passthrough:
        - { source_path: segment_id }
        - { source_path: bbox }
        - { source_path: detection_score }
        - { source_path: quality_score }
  - name: keyframes
    source: { type: collection, collection_id: video-segments }
    feature_extractor:
      feature_extractor_name: image_extractor
      version: v1
      input_mappings:
        image: content
      field_passthrough:
        - { source_path: segment_id }
        - { source_path: frame_ts }
  - name: onscreen-text
    source: { type: collection, collection_id: video-segments }
    feature_extractor:
      feature_extractor_name: scrolling_text_extractor
      version: v1
      input_mappings:
        video: content
      field_passthrough:
        - { source_path: segment_id }
  - name: age-estimation
    source: { type: collection, collection_id: faces }
    feature_extractor:
      feature_extractor_name: age_estimator  # template-provided extractor, installed with the template
      version: v1
      input_mappings:
        face: content
      field_passthrough:
        - { source_path: person_id }
        - { source_path: age_median }
        - { source_path: aged }
        - { source_path: n_looks }
  - name: violence-context
    source: { type: collection, collection_id: keyframes }
    feature_extractor:
      feature_extractor_name: context_scorer  # template-provided extractor, installed with the template
      version: v1
      input_mappings:
        image: content
      field_passthrough:
        - { source_path: segment_id }
        - { source_path: staged_score }
        - { source_path: real_score }
        - { source_path: margin }
clean_views:  # retriever_transform collections, the downstream contract
  - name: person-verdicts
    source: { type: collection, collection_ids: [faces, age-estimation] }
    retriever_transform:
      unique_id_field: person_id
      write_back_fields:
        - { source_field: age_median, target_field: age_median }
        - { source_field: aged, target_field: aged }
        - { source_field: n_looks, target_field: n_looks }
  - name: segment-verdicts
    source: { type: collection, collection_ids: [person-verdicts, violence-context, onscreen-text, video-segments] }  # via retriever pipeline
    retriever_transform:
      retriever: roll-up-by-segment  # the pipeline defines transform + doc identity
      unique_id_field: segment_id
      write_back_fields:
        - { source_field: decision, target_field: decision }
        - { source_field: reason, target_field: reason }
        - { source_field: n_person, target_field: n_person }
        - { source_field: n_minor, target_field: n_minor }
        - { source_field: n_unaged, target_field: n_unaged }
        - { source_field: n_borderline, target_field: n_borderline }
retrievers:
  - name: brand-mark-match
    collections: [brand-mark-index]
    feature_search:
      fusion: rrf
      top_k: 5
      searches:
        - { feature_uri: "mixpeek://image_extractor@v1/google_siglip_base_v1", query: { input_mode: text, text: "{{INPUT.query}}" } }
  - name: public-figure-match
    collections: [person-ref-index]
    feature_search:
      fusion: rrf
      top_k: 5
      searches:
        - { feature_uri: "mixpeek://face_identity_extractor@v1/insightface__arcface", query: { input_mode: text, text: "{{INPUT.query}}" } }
  - name: roll-up-by-segment
    collections: [person-verdicts, violence-context, onscreen-text, video-segments]
    stages:  # pipeline retriever: groups documents, materialises into the collection on its out-edge
      - stage_name: group
        config: { stage_id: group_by, parameters: { group_by_field: segment_id } }
  - name: review-queue
    collections: [segment-verdicts]
    feature_search:
      fusion: rrf
      top_k: 100
      searches:
        - { feature_uri: "mixpeek://multimodal_extractor@v2/gemini-embedding-2", query: { input_mode: text, text: "{{INPUT.query}}" } }
  - name: moderation-search
    collections: [segment-verdicts]
    feature_search:
      fusion: learned
      top_k: 25
      searches:
        - { feature_uri: "mixpeek://multimodal_extractor@v2/gemini-embedding-2", query: { input_mode: text, text: "{{INPUT.query}}" } }
        - { feature_uri: "mixpeek://multimodal_extractor@v2/multilingual_e5_large_instruct_v1", query: { input_mode: text, text: "{{INPUT.query}}" } }
        - { feature_uri: "mixpeek://multimodal_extractor@v2/multilingual_e5_large_instruct_ocr_v1", query: { input_mode: text, text: "{{INPUT.query}}" } }
      learning_config: { shadow_mode: true, rollout_pct: 0.0, context_features: [INPUT.session_id] }  # flip via retriever PATCH when the signals justify it
  - name: cleared-set-export
    collections: [segment-verdicts]
    feature_search:
      fusion: rrf
      top_k: 1000
      searches:
        - { feature_uri: "mixpeek://multimodal_extractor@v2/gemini-embedding-2", query: { input_mode: text, text: "{{INPUT.query}}" } }