Hair Tone Estimation at Roots via Imaging Device with Embedded Deep Learning
January 2019
in “
Electronic Imaging
”
Studysummary This study found that a lightweight Convolutional Neural Network model can accurately and quickly determine natural hair tone from high-resolution images of hair roots, outperforming other popular methods.
Our plain-language summary. Not medical advice or a treatment recommendation. Consult a qualified healthcare professional before changing treatment. Full disclaimer
The study focused on developing a device that uses a Convolutional Neural Network (CNN) to accurately estimate natural hair tone at the roots, which are unaffected by dyeing or environmental conditions. The device captures high-resolution images of hair roots and processes them using a CNN trained on a dataset evaluated by color experts. The proposed model demonstrated higher precision and faster computation times compared to other CNNs and conventional image processing methods, achieving real-time results on a low-end chip. This advancement had significant implications for hair coloration, beauty personalization, and clinical evaluation.
