Deep Learning Based Face Recognition System For Missing Children Identification

Abstract

Thousands of children are reported missing in India every year, and a significant proportion of these cases remain unresolved due to the difficulty of identifying children across large populations. Traditional search methods are time-consuming and often ineffective when dealing with large image datasets. To address this challenge, this study proposes an automated missing child identification system using deep learning–based facial recognition.The proposed system enables the public to upload photographs of suspected missing children through a centralized online portal along with location information and remarks. The uploaded images are automatically compared with images stored in a database of reported missing children. The system then identifies the most probable match by analyzing facial features extracted from the images.A Convolutional Neural Network (CNN) based approach is utilized to perform facial feature extraction. Specifically, a pre-trained VGG-Face deep architecture is employed to generate robust face descriptors from input images. Unlike conventional deep learning pipelines that rely solely on CNNs for classification, the proposed framework uses the CNN model primarily as a feature extractor. The extracted facial embeddings are then classified using a Support Vector Machine (SVM) classifier to determine the identity of the child.The use of VGG-Face enables the system to generate highly discriminative facial representations that remain robust against variations such as lighting conditions, image noise, facial pose, occlusions, and gradual facial changes due to aging. Experimental evaluation was conducted on a dataset consisting of 43 missing child cases. The results demonstrate that the proposed method achieves a classification accuracy of 99.41%, outperforming several traditional face recognition techniques for missing child identification.The proposed approach provides an efficient and scalable solution for assisting law enforcement agencies and the public in locating missing children, thereby improving the likelihood of successful recovery.

DOI: https://doi-ds.org/doilink/04.2026-69328734