What are MediaPipe Face Landmarker face blendshapes and how are they used to measure facial expressions in web applications?
10/2/2026, 4:28:31 PM · llm:deepseek:deepseek-v4-flash + heuristic (fallback from llm:cloudflare:@cf/meta/llama-3.3-70b-instruct-fp8-fast) (fallback from llm:mimo:mimo-v2.5) on 1 step
> ⚠ Low confidence — 1 sub-claim remains below the evidence threshold within budget. Treat this as provisional.
What are MediaPipe Face Landmarker face blendshapes?
The supplied source describes face blendshapes only briefly. It states that face blendshapes are an optional output controlled by the output_face_blendshapes setting, and that they are used for rendering the 3D face model . The source does not provide a fuller definition of what blendshapes are (e.g., a list of named coefficients or their semantic meaning), so that part of the question remains unanswered by the provided passages.
How are they used to measure facial expressions in web applications?
The source indicates that the Face Landmarker task can detect facial expressions and that a third model uses landmarks to identify facial features and expressions . It also notes that platform-specific guides, including a Web code example guide, walk through a basic implementation of the task . However, the supplied passages do not describe the specific procedure for using blendshapes to measure facial expressions in a web application (e.g., how blendshape values are read, interpreted, or mapped to expressions). That part of the question is therefore not answered by the provided sources.
Gaps: The passages do not define blendshapes beyond their rendering purpose, and do not explain the web-application workflow for measuring expressions with them.
Evidence ledger — supporting quotes
What are MediaPipe Face Landmarker face blendshapes?
40%“Float [0.0,1.0] 0.5 output_face_blendshapes Whether Face Landmarker outputs face blendshapes.” [S1] Face landmark detection guide
“Face blendshapes are used for rendering the 3D face model.” [S1] Face landmark detection guide
How are MediaPipe Face Landmarker face blendshapes used to measure facial expressions in web applications?
0%No supporting evidence
Research evidence matrix
Compare research claims with cited sources and inspect recorded excerpts. An empty cell means no inspectable excerpt was recorded; it does not establish whether a claim is true, false, or disputed. Coverage and agent confidence are not measured accuracy.
| Research claim | Inspection status | [S1] Face landmark detection guidePublication: google.comPublished: Not recorded |
|---|---|---|
| What are MediaPipe Face Landmarker face blendshapes? | Recorded excerpt | Inspect 2 excerptsFloat [0.0,1.0] 0.5 output_face_blendshapes Whether Face Landmarker outputs face blendshapes. Face blendshapes are used for rendering the 3D face model. |
| How are MediaPipe Face Landmarker face blendshapes used to measure facial expressions in web applications? | No inspectable excerpt recorded | No excerpt recorded |
Reference export
1 article references. Recorded titles, links and dates; observed scholarly records also include supplied authors, DOI and journal metadata with read limits. Review metadata before using in a paper. Import RIS into Zotero with File → Import.
Cited sources and references
- 1Face landmark detection guidegoogle.comFree public reference · no creator payment · extracted html text100%
Decision log · 70 steps
Breaking down: "What are MediaPipe Face Landmarker face blendshapes and how are they used to measure facial expressions in web applications?"
Identified 2 research target(s) to investigate; these are not established facts
Quick mode: at most 2 paid/cached/public reads, with no marketplace probe or gap-expansion round.
Web search: 2/2 planned queries attempted, 2 succeeded, 18 public page previews, 0 unavailable queries. Snippets are discovery only. Public reads spend no USDC; model and service operating costs remain separate.
Discovered 21 verified creator source(s) and 23 free public reference(s)
Recalled 11 past runs on this subject — how these sources performed when they were available.
ERC-8004 reputation loaded — composite scores on this subject.
Claim-aware portfolio (exhaustive; bounded selection, not a claim of global optimality) selected 2/6 positive proposal(s): 2 free/cache selections + 0 paid fresh selections, predicting 2/2 claim(s) above the evidence floor with $0.000000/$0.050000 fetch USDC reserved.
Free-preview pre-check maps an actionable source to every sub-claim (2/2); paid reading may proceed within the budget.
Strong topical match on mediapipe, face, landmarker, blendshapes, facial, addresses sub-claim 1 & 2; worth the 0 USDC toll. - free public original-page READ selection (not a cache hit); no purchase or creator reward. — selected for the claim-aware evidence portfolio (targets claims 1, 2; 0 fetch USDC, 1 attention slot).
Strong topical match on mediapipe, face, landmarker, facial, expressions, addresses sub-claim 1 & 2; worth the 0 USDC toll. - free public original-page READ selection (not a cache hit); no purchase or creator reward. — selected for the claim-aware evidence portfolio (targets claims 1, 2; 0 fetch USDC, 1 attention slot).
Already cached and still relevant (matches applications); reuse for free instead of paying again. - free public feed reference; no purchase or creator reward. — the free-preview coverage check could not connect this source to any sub-claim, so no toll is authorized.
Already cached and still relevant (matches applications); reuse for free instead of paying again. - free public feed reference; no purchase or creator reward. — the free-preview coverage check could not connect this source to any sub-claim, so no toll is authorized.
Weak match (no key terms); not worth 0 USDC. - free public feed reference; no purchase or creator reward.
Already cached and still relevant (matches they); reuse for free instead of paying again. - free public feed reference; no purchase or creator reward. — the free-preview coverage check could not connect this source to any sub-claim, so no toll is authorized.
Weak match (no key terms); not worth 0 USDC. - free public feed reference; no purchase or creator reward.
Weak match (only web); not worth 0 USDC. - free public original-page READ selection (not a cache hit); no purchase or creator reward.
Weak match (only web); not worth 0 USDC. - free public original-page READ selection (not a cache hit); no purchase or creator reward.
Weak match (only web); not worth 0 USDC. - free public original-page READ selection (not a cache hit); no purchase or creator reward.
Weak match (only web); not worth 0 USDC. - free public original-page READ selection (not a cache hit); no purchase or creator reward.
Weak match (only web); not worth 0 USDC. - free public original-page READ selection (not a cache hit); no purchase or creator reward.
Weak match (only web); not worth 0 USDC. - free public original-page READ selection (not a cache hit); no purchase or creator reward.
Weak match (only web); not worth 0 USDC. - free public original-page READ selection (not a cache hit); no purchase or creator reward.
Weak match (only web); not worth 0 USDC. - free public original-page READ selection (not a cache hit); no purchase or creator reward.
Weak match (only web); not worth 0 USDC. - free public original-page READ selection (not a cache hit); no purchase or creator reward.
Weak match (only web); not worth 0 USDC. - free public original-page READ selection (not a cache hit); no purchase or creator reward.
Strong topical match on mediapipe, face, landmarker, facial, expressions, addresses sub-claim 1 & 2; worth the 0 USDC toll. - free public original-page READ selection (not a cache hit); no purchase or creator reward. — the claim-aware portfolio chose a stronger, less redundant set inside the 2-source attention and $0.050000 fetch-budget caps, so this proposal stays unspent.
Strong topical match on mediapipe, face, landmarker, blendshapes, facial, addresses sub-claim 1 & 2; worth the 0 USDC toll. - free public original-page READ selection (not a cache hit); no purchase or creator reward. — the claim-aware portfolio chose a stronger, less redundant set inside the 2-source attention and $0.050000 fetch-budget caps, so this proposal stays unspent.
Strong topical match on mediapipe, face, facial, web, addresses sub-claim 1 & 2; worth the 0 USDC toll. - free public original-page READ selection (not a cache hit); no purchase or creator reward. — cached bytes are free, but this read does not clear the attention gate (EV 0.36, minimum 0.45, with a required claim target).
Strong topical match on mediapipe, face, landmarker, web, addresses sub-claim 1 & 2; worth the 0 USDC toll. - free public original-page READ selection (not a cache hit); no purchase or creator reward. — cached bytes are free, but this read does not clear the attention gate (EV 0.36, minimum 0.45, with a required claim target).
Strong topical match on mediapipe, face, blendshapes, facial, web, addresses sub-claim 1 & 2; worth the 0 USDC toll. - free public original-page READ selection (not a cache hit); no purchase or creator reward. — the claim-aware portfolio chose a stronger, less redundant set inside the 2-source attention and $0.050000 fetch-budget caps, so this proposal stays unspent.
Strong topical match on mediapipe, face, web, addresses sub-claim 1 & 2; worth the 0 USDC toll. - free public original-page READ selection (not a cache hit); no purchase or creator reward. — cached bytes are free, but this read does not clear the attention gate (EV 0.27, minimum 0.45, with a required claim target).
Weak match (no key terms); not worth 0.003 USDC.
Weak match (no key terms); not worth 0.004 USDC.
Weak match (no key terms); not worth 0.005 USDC.
Weak match (no key terms); not worth 0.003 USDC.
Weak match (no key terms); not worth 0.002 USDC.
Weak match (no key terms); not worth 0.002 USDC.
Weak match (no key terms); not worth 0.002 USDC.
Weak match (no key terms); not worth 0.002 USDC.
Already cached and still relevant (matches they); reuse for free instead of paying again. — the free-preview coverage check could not connect this source to any sub-claim, so no toll is authorized.
Already cached and still relevant (matches web); reuse for free instead of paying again. — the free-preview coverage check could not connect this source to any sub-claim, so no toll is authorized.
Weak match (no key terms); not worth 0.003 USDC.
Weak match (only face); not worth 0.003 USDC.
Weak match (no key terms); not worth 0.004 USDC.
Already cached and still relevant (matches they); reuse for free instead of paying again. — the free-preview coverage check could not connect this source to any sub-claim, so no toll is authorized.
Strong topical match on used, web, addresses sub-claim 2; worth the 0.002 USDC toll. — the claim-aware portfolio chose a stronger, less redundant set inside the 2-source attention and $0.050000 fetch-budget caps, so this proposal stays unspent.
Already cached and still relevant (matches they, web); reuse for free instead of paying again. — the free-preview coverage check could not connect this source to any sub-claim, so no toll is authorized.
Weak match (only they); not worth 0.002 USDC.
Weak match (only web); not worth 0.002 USDC.
Weak match (no key terms); not worth 0.003 USDC.
Already cached and still relevant (matches web); reuse for free instead of paying again. — the free-preview coverage check could not connect this source to any sub-claim, so no toll is authorized.
Weak match (no key terms); not worth 0.002 USDC.
READ MediaPipe Blendshapes recording and filtering - selected original public page, 0 USDC; not a cache hit.
Public page unavailable (transport-unavailable); no evidence admitted. Continuing research.
READ Face landmark detection guide | Google AI Edge | Google for Developers - selected original public page, 0 USDC; not a cache hit.
Read extracted public text from https://developers.google.com/edge/mediapipe/solutions/vision/face_landmarker - S1; quote matching establishes source grounding, not fact verification.
Final check — "What are MediaPipe Face Landmarker face blendshapes?": 40% assessed by S1
Final check — "How are MediaPipe Face Landmarker face blendshapes used to m…": 10% assessed by S1
Final coverage assessment — The supplied excerpt defines face blendshapes only minimally as outputs used for rendering a 3D face model and notes that a third model identifies facial features and expressions. It does not explain what blendshapes are in detail, nor does it describe how they are used to measure facial expressions in web applications beyond a generic Web code example guide link. The assessment does not establish a complete supported answer for every requested part.
Synthesizing a grounded answer from 1 source(s)…
Relevance review returned; only checked excerpts can retain support, and review cannot raise it.
Verified public reference (no creator reward) — S1 supports claim 1 at 50%: “Float [0.0,1.0] 0.5 output_face_blendshapes Whether Face Landmarker outputs face blendshapes.”
Verified public reference (no creator reward) — S1 supports claim 1 at 40%: “Face blendshapes are used for rendering the 3D face model.”
Below support/reward gate — S1 supports claim 2 at 20%: “The first model detects faces, a second model locates landmarks on the detected faces, and a third model uses those landmarks to identify fa…”
Below support/reward gate — S1 supports claim 2 at 10%: “Guide Web Code example Guide Task details This section describes the capabilities, inputs, outputs, and configuration options of this task.”
Drafted answer citing 1 source(s)
Confidence: Low — 1 sub-claim remains below the evidence threshold.
google.com contributed 100% - free public reference; reward share withheld
Done. Spent $0 across 0 confirmed/simulated payment(s) to creators.
Portable research receipt
Take the evidence trail with you
One deterministic JSON bundle binds the answer, visible decisions, exact article versions, claim evidence and a Circle-settlement snapshot under SHA-256. Retain the digest to detect later changes; the self-check is not a publisher or Keryx signature.
Carries this dispatch’s question as context — never its answer. The next dispatch is read from sources bought for it.