Image Processing Based Attendance System with Dual-Stage LSTM and Machine Learning Models

Introduction: The researchers introduced an Image Processing Based Attendance System with Dual-Stage LSTM and Machine Learning Models a pioneering solution in automated attendance tracking. This system innovatively combines the precision of facial landmark detection with the advanced learning capabilities of a Dual-Stage Long Short-Term Memory (LSTM) network. Objectives: The Advanced Attendance Management System with Dual-Stage LSTM and Facial Landmark Detection represents a comprehensive approach to automating attendance tracking through state-of-the-art facial recognition technologies. This research study delineates the system's architecture, detailing each module's function within the framework. The objective function used in this work is Sparse categorical Cross Entropy. The accuracy measure is calculated and reported at each epoch, and the weights are set using the efficient ADAM optimization technique. Methods: This research study is designed to address the complexities of real-world environments, the system sets a new benchmark in recognizing and tracking individual faces over time. The Dual-Stage LSTM for face recognition operates by extracting and analyzing immediate facial features to identify distinct characteristics in the short term. Subsequently, the second stage processes these features over extended periods, learning and adapting to temporal variations in appearance, ensuring accurate long-term recognition despite changes in facial attributes. Results: In implementing this system, special attention has been given to enhancing accuracy and reducing false positives, critical parameters in any attendance management application. By processing facial data through this dual-stage approach, the system demonstrates remarkable proficiency in handling variability in lighting, orientation, and background conditions. Conclusions: The Advanced Attendance Management System, leveraging Dual-Stage LSTM and Facial Landmark Detection, represents a breakthrough in automating attendance with cutting-edge facial recognition technology. Beginning with high-resolution image capture in the Input Module, the system meticulously processes these images through stages including pre-processing for image refinement, facial detection using advanced algorithms like the Haar Cascade Classifier and Viola-Jones, and precise feature extraction via the dlib library.