• Arshi Afreen

The present study introduced an innovative method for recognising masked faces through the utilisation of Latent Part Detection (LPD). The method under consideration employs a convolutional neural network (CNN) for the purpose of identifying masked faces. The identified faces are subsequently processed by a secondary network to produce an embedding for the masked face. The process of utilising this embedding involves comparing the facial features of a masked individual with a pre-existing database of recognised individuals. Comparing the proposed approach to other facial recognition techniques, the findings demonstrate that it achieves high accuracy. Furthermore, the VGG-16 model under consideration has undergone training not only with the ALexNet architecture but also with the ResNet architecture, in conjunction with data obtained from train sets. The present study employed Alex and Res Net for the purpose of evaluating and contrasting the efficacy of the proposed model with respect to the performance metrics. The results indicated a 97% accuracy rate, 97.62% specificity rate, and 100% sensitivity rate, which surpasses the performance of previous models. The article additionally offers valuable perspectives on the characteristics of the data and elucidates the potential applications of LPD in enhancing the precision of facial recognition. Our aim in the forthcoming period is to broaden the scope of our masked dataset to encompass supplementary applications that involve masked faces. Simultaneously, we endeavour to consistently augment the efficacy of masked face recognition. Keywords: Masked face, LPD, biometric technology, CNN.

Masked face, LPD, biometric technology, CNN

"MASKED FACE RECOGNITION WITH LATENT PART DETECTION", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.12, Issue 1, page no.c393-c404, January-2025, Available :http://www.jetir.org/papers/JETIR2501247.pdf

"MASKED FACE RECOGNITION WITH LATENT PART DETECTION", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.12, Issue 1, page no. ppc393-c404, January-2025, Available at : http://www.jetir.org/papers/JETIR2501247.pdf