Draw-a-Deep Pattern: Drawing Pattern-Based Smartphone User Authentication Based on Temporal Convolutional Neural Network

Present-day smartphones provide various conveniences, owing to high-end hardware specifications and advanced network technology. Consequently, people rely heavily on for a myriad of daily-life tasks, such as work scheduling, financial transactions, social networking, which require strong robust user authentication mechanism protect personal data privacy. In this study, we propose draw-a-deep-pattern (DDP)—a deep learning-based end-to-end smartphone method using sequential obtained from drawing character or freestyle pattern the touchscreen. our model, recurrent neural (RNN) temporal convolution (TCN), both are specialized in processing, employed. The main advantages proposed DDP (1) it is threats current systems vulnerable, e.g., shoulder surfing attack smudge attack, (2) requires few parameters training; therefore, model can be consistently updated real-time, whenever new training available. To verify performance collected 40 participants one most unfavorable environments possible, wherein all potential intruders know how authorized users draw characters symbols (shape, direction, stroke, etc.) used authentication. Of two models, TCN-based yielded excellent with average values 0.99%, 1.41%, 1.23% terms AUROC, FAR, FRR, respectively. Furthermore, exhibited improved higher computational efficiency than RNN-based cases. contribute research/industrial communities, made dataset publicly available, thereby allowing anyone studying developing behavioral biometric-based system use without any restrictions.