Humanity Protocol (w Tokenomics)
  • Introduction to Humanity Protocol
    • We Are Solving the Identity Problem
    • The $H Token
  • Human-Centric Blockchain
    • Why Does Humanity Protocol Matter
    • Unlocking New User Cases
    • How Does Proof of Humanity Work
    • Key Players and Components of the HP Ecosystem
      • Human Recognition Module
      • Unique Human Users
      • Privacy-Preserving Data Storage and Use
      • Identity Validators
      • zkProofer Nodes
      • Proof of Humanity (PoH) User Journey
      • Product Development and Privacy Roadmap
  • HP Software and Hardware DePIN Network
    • Why Palm Recognition
    • Humanity Palm Recognition AI Model
    • Initial Phase: Advanced Palmprint Recognition
    • Second Phase: DePIN of Humanity Scanners
  • Tokenomics
    • Humanity Protocol Ecosystem & Stakeholders
    • Token Allocations
    • Token Lockups and Emissions
    • Identity Validator and Staking Rewards
    • Risks and Disclosures
  • zkProofer Node Distribution
    • Distribution Process
    • Node Incentive Mechanism
Powered by GitBook
On this page
  1. HP Software and Hardware DePIN Network

Humanity Palm Recognition AI Model

Humanity Protocol leverages advanced biometric technology and a visionary approach to decentralized networks to redefine digital identity verification. Central to this transformation is HP's specialized hardware, designed not only for state-of-the-art palm vein recognition but also for pivotal roles in the protocol’s future DePIN networks.

Humanity Protocol continues to develop a proprietary AI model, designed to significantly improve the precision of palm recognition. This model leverages deep learning algorithms, enhancing the processing of palm images by learning from a vast, encrypted database of palm features.

Our proprietary convolutional neural network (CNN) architectures for palm print and palm vein detection and identification significantly surpass traditional frameworks such as ResNet18 in terms of model generalization and compression efficiency. Engineered for deployment on prevalent SoC chips with 1T computational capacity, these models achieve an impressive footprint of less than 2MB. This optimization facilitates a streamlined pipeline—from image ingestion, palm detection and feature extraction, to final output generation—all within less than 0.1 seconds.

PreviousWhy Palm RecognitionNextInitial Phase: Advanced Palmprint Recognition