Security framework based on Biometric recognition is highly acknowledged in the research community. However, security challenges have been seen in the existing research related to recognition failure and security issues. The proposed model endorses a paradigm of biometric recognition where the focus is to maintain the perseverance of the recognition accuracy under any unpredictable circumstances. The model utilizes a set of nonsymmetrical dynamic video inputs in which facial subjects have been traced and analyzed. The aim of the work is to provide an effective and consistent recognition rate and preserve the security of the biometric characteristics. The proposed model uses the Haar-Cascade algorithm to trace the facial subject from an input video frame. This algorithm utilized the haar-plotting method to obtain the facial portion accurately. Then, the model applied preprocessing operations to normalize the variation and noise.