Harnessing Artificial Intelligence at Frontier.cool to Enhance the Accuracy of Fabric Physical Properties
Introduction
The physical properties of Frontier.cool's digital fabrics are measured using prediction models powered by the Lasagna AI Engine. It analyzes fabric layers from 2D images to generate 3D texture maps, leveraging extensive material data like type, weight, and features. The model’s performance is validated by comparing predicted properties with real-world fabric behavior from physical tests or simulations, ensuring accuracy and reliability.
Input Variables
The input variables consist of (1) material type, (2) construction type, (3) content, and (4) weight. The detail of each variable is explained:
1. Material Type
There are five material types: woven, knit, leather, non-woven, trims, and others.
2. Construction Type
The construction type is based on the material type, for example, knit, denim, canvas, challis, mesh, lace, satin, twill, etc.
3. Content
The content consists of one fiber or more than one fiber. The fibers include acrylic, amicor, biofibre, nylon, chiffon, cotton, polyester, rayon, silk, tencel, vinyl, wool, etc.
4. Weight (Mass Density)
The default weight is in g/m2, with a minimum weight of 1.00 g/m2, while the other units are oz/yd2, g/y, and g/m.
Output Variables
Each fabric sometimes uses different variables and naming conventions, for example, the general terms "length" and "width" to describe fabric size and properties, and another one uses "warp (vertical)" and "weft (horizontal)." These terms refer to the specific directions in which yarns are woven, reflecting the structural differences of fabrics in their detailed simulations.
Variables
- Bending (length and width or warp and weft): This measures a fabric's stiffness or flexibility, with values ranging from 0 to 1000. A value of 0 indicates no rigidity, 5 is low flexibility, and 1000 signifies high stiffness. Low values mean the fabric is soft and bends easily, while high values indicate stiffness, like leather.
- Bending Bias: This refers to the directional bending behavior of fabric, affecting asymmetrical effects in garments. For practical purposes, assuming symmetrical bending, the Bending Bias is the average of Bending Warp and Bending Weft values.
- Stretch (Tensile): This indicates how much a fabric can be extended or its elasticity under force.
- Stretch Linearity: This evaluates how evenly a fabric stretches, expressed as a percentage. High linearity means the fabric stretches consistently, while low linearity indicates it becomes difficult to stretch quickly, as seen in woven fabrics.
- Shear: This measures the force needed to distort or shift fabric at a 45-degree angle to its grain.
- Thickness (Compression): This describes how thick or thin a fabric feels, measured in millimeters. Thickness affects a fabric's appearance, drape, and performance. Thicker fabrics tend to be heavier, more durable, and better at insulation due to trapped air.