Ai-driven fit optimization for absorbent underwear using 3D body scans

Overview

Baek Labs, LLC presents an innovative AI-driven platform designed to optimize the fit of absorbent underwear products. This solution employs generative AI to create realistic 3D models of human pelvic geometries and integrates a physics-aware machine learning system to analyze product designs. The aim is to enhance comfort and movability by tailoring product designs to a diverse range of body shapes, leveraging existing data from over 4000 3D body scans.

Technical specifications

• Generative AI: Generates a wide array of anthropometrically accurate 3D body models using existing 3D scan data.
• AI-accelerated physics solver: Rapidly evaluates physical properties such as pressure, temperature, and permeability on product designs.
• Graphical neural networks: Learn from 3D body scan data to produce realistic 3D geometry models.
• Finite element mesh (FEM) models: Used to analyze fit and physical interactions between body and product.
• Interactive design tool: Provides real-time fit analysis, enabling quick design iterations and optimizations.

Technology readiness level

Currently at TRL 5, this technology has been validated in a relevant environment. The platform's components have been tested, and further development will focus on refining the AI models and integrating them into a comprehensive design tool.