ARCHIVES

Secure, reliable, and interpretable user authentication has become critical in the era of pervasive digitalization. Traditional methods such as passwords and token-based logins are increasingly vulnerable to theft and brute-force attacks, while physiological biometrics such as fingerprints or facial recognition raise privacy and spoofing concerns. This study introduces a moving way to check who someone is by how they write - using actions instead of static traits. Instead of fixed passwords, it tracks live movements like grip strength, path shape, speed; these details come from digital pens on tablets. What sets it apart is its mix of smart layers: one part sees shapes in strokes using CNN tech, another catches timing patterns through BiLSTM units working both ways in time. With info gathered from several writers and nearly all data used for training - just 2% held back - it managed correct ID calls more than nine out of ten times during repeated checks. A Streamlit app was built to show live predictions, highlighting how it could work in real use. The findings prove handwritten patterns are strong indicators for future ID verification tech