CRIMINAL FACE DETECTION USING DEEP LEARING

Criminal Face Detection has emerged as a critical technology in enhancing public safety and security. The proposed system encompasses several key components like data collection, data preprocessing, model training, real-time face recognition. Data collection involves assembling a diverse dataset of facial images, ensuring variations in lighting and expression. Data processing includes face detection using Multi-task Cascaded Convolution Network through facial landmarks, normalization and augmentation to enhance datasets variability. Feature extraction employs pre-detriment deep learning model like Face Net generating the robust feature embeddings for each face. The model training focus on the selecting and training the CNN’s Network while addressing the class embrace the utilizing the Over-Sampling Synthetic Minority Technique (SMOTE) In the development the system integrates with the real time surveillance and alerting the criminal face. The alert notification system is implemented to automate the response. Ethical considerations and data privacy are paramount throughout the system development and deployment ensuring compliance with legal standards and protecting individual rights. The proposed system demonstrates significant potential in improving law enforcement capabilities and enhancing public safety through accurate and efficient criminal face detection.