Assessment of Safety and Privacy Challenges in ANN-Based English Handwriting Recognition

Deep neural network-based English handwriting recognition has revolutionised the security or verification system of today; nevertheless, there are certain risks that are inevitable.This paper examines the status of the present technologies in the handwriting recognition systems, and more particularly, by the various deep neural network architectures.It also evaluated common cybersecurity risks such as data poisoning, model inversion, and adversarial attacks, which can be devastating to such systems, as well as common privacy and ethical issues.The potential regulatory compliance and mitigation measures that can be taken to avert these risks and hurdles are also addressed in detail, with requisite emphasis being made on the future outlook of a more secure handwriting recognition system.