Skip to content

πŸš€ HyperX β€” Web3 Medical Image Labeling with Transparent Rewards

πŸ—‚οΈ Project Overview

  • Student ID: 20240310
  • Name: Jimin Park
  • Project Title: HyperX β€” Web3 Medical Image Labeling with Transparent Rewards
  • Summary (3–4 sentences): HyperX is a Web3 platform where medical experts label ultrasound/CT/MRI images and receive transparent, algorithmic rewards. We combine a tier system with base pay, performance pay, and contribution bonuses; add meta-review for disagreements; and use verification items, consistency checks, and invisible watermarks for trust and safety. Users log in with a wallet and see staking, rewards, and pool share on a dashboard. A Zero-Cost mode lets anyone experience the full flow without gas, while Anvil mode runs the same dApp on a local chain.

1. 🧩 Problem: What Problem Are You Solving?

  • Quality medical labels are scarce and slow to produce; traditional outsourcing has opaque quality, timelines, and incentives.
  • There’s no standardized, auditable way to translate expert effort and difficulty into fair rewards.
  • Enterprises need provable trust (QC, audit trails) and strong data-leak safeguards.

2. πŸ’‘ Solution: Your Proposed Approach

  • Core design
    • Reward mechanics: Base pay (tier & volume brackets) + performance pay (correctness Γ— tier weight) + contribution bonus for β€œhard” items (accuracy-rate inverse Γ— base). Final payouts come proportionally from a Reward Pool.
    • Trust stack: Verification items (easy, repeatable), consistency scoring, meta-review when agreement < 90%, stake/escrow to deter careless work, and invisible watermarking to prevent leaks.
    • Task routing: Rules/cluster-based recommendations (later RL/bandits) to match experts to tasks by skill and history.
  • UX
    • Wallet login β†’ How it works explainer β†’ Label tasks (public demo images) β†’ Consensus/meta-review β†’ Claim β†’ Dashboard shows rewards, stake, and pool share.
  • (Optional) Include a diagram to show the system architecture or workflow. CleanShot_2025-07-24_at_11.09.062x CleanShot_2025-07-24_at_11.16.212x 1

3. πŸ”— Why Blockchain (and Token)?

  • Transparent settlement: Immutable, queryable records for submissions, agreement, and payouts.
  • Aligned incentives: On-chain reward pool, staking/escrow, and tiered economics minimize ambiguity and disputes.
  • Token (future):
    • Utility/reward token for payouts & staking.
    • Governance token for voting on rules (tiers, thresholds, pool splits).
  • Cost control today: Zero-Cost simulation (off-chain) for demos; Anvil mode for local on-chain testing.

4. πŸ› οΈ MVP or Prototype

  • Current status: β˜‘ Working MVP (Zero-Cost & Anvil modes) ☐ Idea only ☐ Prototype ☐ Working MVP
  • Key features you implemented (1–2 short paragraphs)
    • Solidity contracts: LabelingHub (register/stake/consensus/meta-review/claim), MockUSDC (test token), Reward Pool split (base/performance/contribution).
    • Frontend (Next.js + wagmi/viem): Wallet login, How it works page, task labeling with web-fetched demo images, dashboard for rewards/stake/share, Faucet and Claim.
    • Zero-Cost mode: Full off-chain simulation, identical UX with no gas.
    • Testing: Foundry unit/property tests for economics; Playwright E2E for wallet-mock β†’ label β†’ claim.
  • Code repository: Github
  • Screenshots or demo images:
    (Attach 2–3 images if available) CleanShot 2025-08-13 at 09.05.59@2x CleanShot 2025-08-13 at 09.06.08@2x CleanShot 2025-08-13 at 09.06.13@2x CleanShot 2025-08-13 at 09.06.21@2x

5. πŸ“¬ Submission to Hackathons or Grant Programs

  • Name of the program you submitted to: CoinFabric
  • Submission link: CoinFabric
  • Short summary of your submitted description: A Web3 labeling MVP that encodes trust and fair rewards in code, with Zero-Cost demo flow and local on-chain mode.
  • (Optional) Screenshot of the submission confirmation or page CleanShot 2025-08-13 at 08.53.57@2x CleanShot 2025-08-13 at 08.54.29@2x
  • Name of the program you submitted to: ICISTS GRAFFITI 2025(only idea)
  • Short summary of your submitted description: Submitted Idea on ICISTS GRAFFITI 2025(4-days ideathon)
  • (Optional) Screenshot of the submission confirmation or page CleanShot 2025-08-13 at 09.14.31@2x You can see more details on Presentation File

6. πŸ€” Reflection & Future Work

  • What did you learn during this project?

    • How to encode incentive design (base pay, performance pay, contribution bonus, staking) directly in smart contracts and reflect it cleanly in the UI.
    • Practical Web3 product lessons: wallet UX friction, the value of a Zero-Cost simulation mode, and keeping sensitive data off-chain while settling rewards on-chain.
    • Using Foundry + Anvil for rapid Solidity iteration and wagmi/viem for a reliable wallet + contract front end.
    • How consensus, meta-review, and verification items can turn ambiguous medical labels into auditable outcomes.
    • Broader understanding of the Web3 ecosystemβ€”trade-offs around cost, transparency, governance, and compliance considerations.
  • What were the biggest challenges?

    • Designing incentives so labelers work diligently and enterprises still find pricing/payouts economically reasonable.
    • Ensuring label quality at scale (consistency checks, verification items, and meta-review thresholds) without creating a heavy UX.
    • Balancing transparency with privacy/security (watermarking, access control, de-identification).
    • Keeping the stack cohesive (contracts ↔ frontend ↔ tests) and maintaining reproducible local runs
  • What would you improve or build next if you had more time?
    • Web2 backend for enterprises: traditional login, DB, and an upload portal so clients can securely submit image sets; full audit logs and access control.
    • A stronger on/off-chain bridge: queue jobs off-chain, commit results/proofs on-chain, and batch settlements to cut costs.
    • Smarter task routing (RL/bandits), richer reviewer tools, and better analytics for label quality and contributor reputation.
    • Privacy hardening: stricter de-identification checks, invisible watermark tuning, and role-based redaction.
    • Gas & UX polish: AA/paymasters, further Zero-Cost simulations, and more E2E tests for wallet flows and claims.

7. πŸ“š References

  • Foundry (Forge/Cast/Anvil), OpenZeppelin, Next.js, wagmi/viem, Playwright
  • Internal docs: docs/PLAN.md, docs/RUNBOOK.md, CLAUDE.md
  • Market notes: Datavant / Doctorworks / MediLabel summaries