Research on color solid built by gridded color mixing of nine primary colored fibers and its neural network color prediction approach

Abstract According to the demand of color prediction for colored yarn, two adjacent colors chosen from red (R), yellow (Y), green (G), cyan (C), blue (B) and magenta (M) fibers were combined with dark gray (O 1 ), medium 2 ) light 3 respectively, then ternary coupling‐superposition mixing was performed acquire a solid consisting three lightnesses, 18 units 18×(m+1)×n grid points. An integrated 20% hue gradient 33.33% saturation achieve containing 360 points using it as sample space model. typical samples established by points, 213 yarns fabrics prepared parameters, corresponding reflectance accessed spectrophotometer. Neural network models predicting ratios well forecasting established. The 12 non‐grid point parameters prepare fabrics, measured. predicted measured values neural model compared verify its predictive ability generalizability. results showed that: when ratios, difference between ranged 1.5 3.4, an average 2.4; color, 0.8 5.6, 2.4. This article is protected copyright. All rights reserved.