Analysis of 3D Trajectory-Based Modeling and Recognition

3D trajectory-based recognition remains a challenging task in public authentication systems. Traditional authentication systems rely on memorized credentials, are susceptible to information leakage, and often depend on fingerprints, making them vulnerable to security breaches. Its capability to categorize attributes such as handedness, gender, and age groups highlights the uniqueness of this feature and its potential for further development in emerging applications. This work presents the utilization of feature extraction and selection techniques for recognition of online characters through the analysis of online data. Recognition has been carried out using various supervised machine learning models using Random Forest, Support Vector Machines, Hidden Markov Model (HMM), and deep learning models like CNN.

Experiments are carried out using performance criteria including accuracy and other parameters. This study demonstrates that ensemble methods such as Random Forest surpass alternative approaches, for early detection and public authentication.