Optimized intelligent strategy for user authorization by facial recognition.

Pamarthi, Pranathi; Lakshmi, C.; Suneetha, M.

The technology of facial expression recognition plays a crucial role in daily life. In many computer-aided fields, the systems are worked based on face authorization. Several models have been developed to recognize the facial expressions. However, the presence of different features created a complexity for the prediction. Also, the model degraded in recognition accuracy. Therefore, a novel Buffalo-based Zfnet Recognition Framework was developed to recognize the face expressions for the user face authorization system. The model works based on preprocessing, feature extraction, and recognition. The image noises are filtered at the preprocessing stage. Further, the facial features are analyzed and recognized using the fitness process of the African Buffalo. The designed Buffalo-based Zfnet Recognition Framework is tested in Python with a face expression recognition dataset. The results regarding error, f1-score, precision, accuracy, and recall are compared with prevailing models. The validated accuracy, recall, f-score, and precision rate of the designed framework is 99.95%, more significant than that of the existing models. Incorporating the African buffalo function in the Zfnet increases prediction accuracy and reduces relative errors. Thus, the recommended recognition system works more effectively than the traditional models.

BUFFALO (N.Y.); AFRICAN buffalo; FEATURE extraction; FACIAL expression

Multimedia Tools & Applications, 2024, Vol 83, Issue 23, p63353

1380-7501

Academic Journal

10.1007/s11042-023-18072-0