A capacitive skin glass for non-contact eye blink detection in complex environments

Abstract

Traditional camera-based blink detection methods often necessitate the incorporation of additional optical components into glasses to mitigate challenges posed by complex lighting conditions, thereby significantly augmenting product complexity and costs. To overcome these obstacles, this study developed a capacitive skin glass for blink detection in intricate environments, relying on capacitance variations instead of visual imagery. The capacitive skin solely requires electrodes and a chip-based circuit, specifically the FDC2214, which facilitate real-time capacitance measurement. In the present work, capacitance values underwent preprocessing via Gaussian weight interpolation and band-pass filtering before integration into the blink detection algorithm. Subsequently, the blink detection results from both eyes were combined. Experimental tests on electrode length and positioning have revealed that an electrode length of 3.5 cm, when attached to the upper edge of glasses frames without obstructing the line of sight, can achieve an 86.1% accuracy in detecting blinks. Comparative experiments with a dlib-based blink detection algorithm indicate that the proposed method exhibits high accuracy and stability in capacitive skin glass under light conditions ranging from 0 to 10 lux. These findings underscore the potentials of the proposed approach for blink detection in low-light environments.