Figure 2 | Scientific Reports

Figure 2

From: Automatic dispersion, defect, curing, and thermal characteristics determination of polymer composites using micro-scale infrared thermography and machine learning algorithm

Figure 2

Thermal image, corresponding predicted, and true image for (a) 2.5% G Planetary, (b) 2.5% G Hand, (c) 10% G Planetary, (d) 10% G Hand. Thermal image is manually fed in image processing to remove artifacts or voids or blurriness (scale bar = 1000 microns, enclosing red circle area remarks artifact due to container mark, enclosing rectangle area remarks void area). The thermal image is then fed into machine learning model to obtain the true fabrication. By feeding a few thermal signatures, this learning method can show true fabrication in a really quick manner.

Back to article page