Enhancing Image Recognition for Security Applications: A Machine Learning Approach

This study investigates machine learning techniques for image recognition in security applications. It employs preprocessing methods like OpenCV, Pillow, and TensorFlow, alongside feature extraction techniques including Wavelet, CNN, and LBP. Various classifiers such as Naïve Bayes, SVM, Logistic Regression, Random Forest, and Decision Tree are evaluated for accuracy prediction, with parameters optimized using Grid Search CV and scaling to prevent overfitting. Among these, Support Vector Machine attains the highest accuracy of 90%, making it the preferred choice for image recognition. The research aims to empower machines in comprehending visual data, crucial for tasks like verifying personnel and identifying suspects in public areas.