You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Make SUBMISSION.md the single source for submission copy
Adds paste-ready elevator pitch and Devpost text description. The
description now opens with the sourced scale of the problem (12 million
custom tailors, PLFS 2023-24) and the 0.5 cm fit-perception finding
before describing features, so the impact case is stated rather than
implied.
Also corrects the Data For India citation to the URL whose content was
actually verified, adds PMID and PMCID to the two research citations,
records the dev.to write-up, adds the evidence ledger to judge access,
and removes all em dashes.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Copy file name to clipboardExpand all lines: SUBMISSION.md
+81-17Lines changed: 81 additions & 17 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -1,26 +1,88 @@
1
1
# PatternProof — submission
2
2
3
-
> Most AI try-on helps one person decide what to buy. PatternProof helps two people agree on what to make — before fabric becomes irreversible.
3
+
> Most AI try-on helps one person decide what to buy. PatternProof helps two people agree on what to make, before fabric becomes irreversible.
4
4
5
-
## Short description
6
-
7
-
PatternProof is a three-key production interlock for made-to-order clothing. YouCam supplies body-specific visual evidence, the tailor supplies the construction promise, and the customer supplies consent. PatternProof binds all three to one frozen SHA-256 Expectation Checksum before fabric is cut.
5
+
This file is the single source for every piece of submission copy. Sections marked **paste-ready** go verbatim into the Devpost form.
8
6
9
7
## Entry details
10
8
11
9
- Hackathon topic: **Apparel VTO**
12
10
- Platform: **Web**
13
11
- Project provenance: **New project built during the 2026 hackathon submission period.**
Most AI try-on helps you decide what to buy. PatternProof helps two people agree on what to make, a feasibility-checked Cut Card, frozen and approved before fabric is cut.
24
+
```
25
+
26
+
---
27
+
28
+
## Devpost text description (paste-ready)
29
+
30
+
> Field: *"Provide a text description explaining the features, functionality, and consumer or retail value of your project."*
31
+
32
+
```
33
+
India alone has around 12 million custom tailors, one in six of all its manufacturing
34
+
workers, 99% of them informal and 72% of them women, per India's Periodic Labour Force
35
+
Survey 2023-24. In much of the world clothing is made, not bought, and that transaction
36
+
has no preview step. A customer holds up a phone and says "like this," the tailor
37
+
interprets, and the fabric is cut either way. Off the rack, a mistake is a return. Made
38
+
to order, it is destroyed cloth that the customer often bought separately, plus unpaid
39
+
rework for the tailor.
40
+
41
+
The industry has been solving measurement. But Ashdown and DeLong (Applied Ergonomics,
42
+
1995, PMID 15677000) found people detect a waist fit difference as small as 0.5 cm and
43
+
differ significantly in what they accept as correct. You cannot resolve a disagreement
44
+
about preference with a better tape measure. You resolve it by showing both people the
45
+
same thing before the cut.
46
+
47
+
PatternProof is a consent-to-cut workflow for made-to-order clothing. Customers or
48
+
tailors can create a private Cut Card without an account, upload a body photo and garment
49
+
reference, confirm image rights and processing consent, and generate body-specific visual
directions, and an optional post-approval motion proof. The result is clearly presented
52
+
as a visual-intent reference, not a guarantee of fit, drape, measurements, or
53
+
construction.
54
+
55
+
The tailor reviews each customer requirement and records whether it can be made as shown,
56
+
requires an adjustment, or is not feasible. Important details can also be pinned directly
57
+
onto the preview. A not-feasible decision blocks customer review. The customer then
58
+
receives a private, expiring link to approve the exact frozen Cut Card or request a
59
+
change. Change requests preserve the previous version and create a new revision instead
60
+
of rewriting history.
61
+
62
+
Cutting is released only when three keys agree: YouCam visual evidence, the tailor's
63
+
construction judgment, and the customer's approval of the same frozen revision. Approved
64
+
Cut Cards include a checksum, revision history, printable QR code, WhatsApp sharing, and
65
+
optional customer-photo erasure. For customers, PatternProof makes expectations and
66
+
adjustments visible before commitment. For tailoring studios and made-to-order retailers,
67
+
it provides a structured, auditable approval process that supports clearer conversations
68
+
and customer trust without replacing professional fitting or craft judgment.
69
+
```
70
+
71
+
---
72
+
73
+
## Short description
74
+
75
+
PatternProof is a three-key production interlock for made-to-order clothing. YouCam supplies body-specific visual evidence, the tailor supplies the construction promise, and the customer supplies consent. PatternProof binds all three to one frozen SHA-256 Expectation Checksum before fabric is cut.
14
76
15
77
## The problem
16
78
17
79
Made-to-order clothing can begin with an imprecise conversation: a customer shows a garment photo and describes what they want, while a tailor interprets it. Fabric may be cut before both people have confirmed the same intended outcome. A tape measure cannot resolve a disagreement about silhouette, coverage, ease, or design preference.
18
80
19
-
This matters at meaningful scale. India has about 12 million custom tailors—more than one in six manufacturing workers—and the cited labour analysis reports that 99% work informally and 72% are women. Peer-reviewed apparel research also shows that people detect small fit differences and differ in what they consider acceptable.
81
+
This matters at meaningful scale. India has about 12 million custom tailors, more than one in six manufacturing workers, and the cited labour analysis reports that 99% work informally and 72% are women. Peer-reviewed apparel research also shows that people detect small fit differences and differ in what they consider acceptable.
20
82
21
-
To test the problem framing—not estimate prevalence—we manually coded 108 eligible 1.0–2.0 public tailoring complaints concerning 41 de-identified businesses across Mumbai, Bengaluru, and Delhi. Expectation mismatch was the primary code for 24/108; eight more allegations concerned damaged customer fabric. That grouping, specified before coding, is 32/108 (29.6%), but this purposive, one-coder negative-review sample is neither product validation nor a claim that PatternProof could have prevented those events. Fit, delay, workmanship, service, and price account for the other 76/108 and remain outside the product promise. The complete method and limitations are in [RESEARCH.md](RESEARCH.md).
83
+
To test the problem framing rather than estimate prevalence, we manually coded 108 eligible 1.0 to 2.0 public tailoring complaints concerning 41 de-identified businesses across Mumbai, Bengaluru, and Delhi. Expectation mismatch was the primary code for 24/108; eight more allegations concerned damaged customer fabric. That grouping, specified before coding, is 32/108 (29.6%), but this purposive, one-coder negative-review sample is neither product validation nor a claim that PatternProof could have prevented those events. Fit, delay, workmanship, service, and price account for the other 76/108 and remain outside the product promise. The complete method and limitations are in [RESEARCH.md](RESEARCH.md).
22
84
23
-
PatternProof does not claim to predict physical fit or eliminate remakes. It creates a verifiable communication checkpoint before an irreversible production step. A pre-results field protocol now fixes the usability tasks, outcome definitions, denominators, privacy limits, and descriptive analysis before any participant data is collected; see [VALIDATION-PROTOCOL.md](VALIDATION-PROTOCOL.md).
85
+
PatternProof does not claim to predict physical fit or eliminate remakes. It creates a verifiable communication checkpoint before an irreversible production step. A pre-results field protocol fixes the usability tasks, outcome definitions, denominators, privacy limits, and descriptive analysis before any participant data is collected; see [VALIDATION-PROTOCOL.md](VALIDATION-PROTOCOL.md).
24
86
25
87
## What it does
26
88
@@ -29,9 +91,9 @@ PatternProof does not claim to predict physical fit or eliminate remakes. It cre
29
91
3. Category-specific capture checks reject images that are too small, dark, or poorly framed before a billable API call.
30
92
4. YouCam Clothes Virtual Try-On V3 generates the core body-specific visual-intent preview. When a reference needs rescue, YouCam Background Removal may run first. After the core preview and before human review, the tailor may choose one YouCam-defined Fabric VTO direction; it is explicitly not an uploaded swatch or drape simulation.
31
93
5. After approval, YouCam Image-to-Video V2 can create a fixed five-second 480p presentation proof. Motion is excluded from the frozen construction checksum and cannot change an approved Cut Card.
32
-
6. The tailor marks every non-negotiable as “Can make as shown,” “Can make with adjustment,” or “Not feasible.” Adjustments require a customer-visible note; a not-feasible item blocks approval.
94
+
6. The tailor marks every non-negotiable as "Can make as shown," "Can make with adjustment," or "Not feasible." Adjustments require a customer-visible note; a not-feasible item blocks approval.
33
95
7. The tailor turns the YouCam result into an Agreement Map: every coordinate-specific pin is linked to one non-negotiable and inherits its green, amber, or blocked decision state.
34
-
8. A shared Cut Readiness Passport exposes six conditions—rights, preview, human decisions, feasibility, frozen snapshot, and customer approval—so “ready to cut” is an explicit protocol rather than a visual impression.
96
+
8. A shared Cut Readiness Passport exposes six conditions (rights, preview, human decisions, feasibility, frozen snapshot, and customer approval) so "ready to cut" is an explicit protocol rather than a visual impression.
35
97
9. Starting customer review freezes the exact preview, reference proof, requirements, decisions, spatial annotations, consent evidence, shop, and revision digest into an Expectation Checksum.
36
98
10. The customer must acknowledge each customer-visible tailor adjustment before approval. The server independently derives the required adjustment IDs from the frozen snapshot and demands an exact match; a blind or bypassed checkbox cannot turn the customer key.
37
99
11. The customer can instead submit a snapshot-bound veto. That request races safely with approval, blocks cutting at the database boundary, and becomes a verified version+1 in Revision Replay rather than silently editing history.
@@ -55,7 +117,7 @@ In our visual review of these three synthetic cases, all three produced recogniz
- Durable provider-task reservations, leases, per-owner throttling, exact costs of 1/2/2/5 units for Background Removal/Clothes/Fabric/Motion, a 12-unit guest ceiling, and one global circuit breaker.
120
+
- Durable provider-task reservations, leases, per-owner throttling, exact costs of 1/2/2/5 units for Background Removal, Clothes, Fabric, and Motion, a 12-unit guest ceiling, and one global circuit breaker.
59
121
- Digest-bound 256-bit customer links with 14-day expiry, revocation, and single-revision approval.
60
122
- Fenced cleanup workers for abandoned intake, revision-clone recovery, and body-photo erasure.
61
123
- No YouCam or Supabase service credential enters browser code. A public evidence ledger makes the bounded transfer tests, private signed-URL path, and observed byte-identical repeat visible without exposing customer or operational data.
@@ -87,21 +149,21 @@ This is not a fitting room, measurement system, tailor POS, marketplace, or fit
87
149
88
150
## Source evidence
89
151
90
-
-[Data For India, “The Rise of Custom Tailoring”](https://www.dataforindia.com/custom-tailoring/) — analysis based on PLFS July 2023–June 2024 and Census projections.
-[Dik et al., psychographic orientation and apparel fit satisfaction](https://pmc.ncbi.nlm.nih.gov/articles/PMC10362334/) — Heliyon, 2023.
152
+
-[Data For India, "The Rise of Custom Tailoring"](https://www.dataforindia.com/the-rise-of-custom-tailoring/), analysis based on PLFS July 2023 to June 2024 and Census projections.
-[Dik et al., psychographic orientation and apparel fit satisfaction](https://pmc.ncbi.nlm.nih.gov/articles/PMC10362334/), Heliyon, 2023, PMCID PMC10362334.
93
155
-[Perfect Corp. AI Clothes Virtual Try-On documentation](https://docs.perfectcorp.com/reference/ai_clothes/section/overview).
94
-
-[PatternProof formative negative-review study](RESEARCH.md) — manual problem discovery; purposive sample, not prevalence or product validation.
156
+
-[PatternProof formative negative-review study](RESEARCH.md), manual problem discovery; purposive sample, not prevalence or product validation.
95
157
96
158
## Judge access
97
159
98
-
- Live URL: [https://patternproof-nu.vercel.app](https://patternproof-nu.vercel.app)
0 commit comments