Hair Tone Estimation at Roots via Imaging Device with Embedded Deep Learning

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.