Authors : Shani Raj; Sabeena K.

Volume/Issue : Volume 8 - 2023, Issue 10 - October

Google Scholar : https://tinyurl.com/y2vbf75r

Scribd : https://tinyurl.com/yjf3ntxz

DOI : https://doi.org/10.5281/zenodo.10071339

Abstract : The issue of juvenile individuals running away or eloping is a serious and ongoing problem that can have significant consequences for the safety and well-being of the juveniles involved, as well as for their families and communities. In this project, we propose a system for detecting and reuniting missing juveniles using artificial intelligence (AI) techniques. The system utilizes data uploaded by the authorities in the juvenile home through a website, in which every juvenile has an individual profile that is used to identify and locate missing juveniles in real-time. By applying HOG feature, Face Landmark Estimation, CNN, and SVM Classifier to this data, we are able to train the model with each juvenile. When the public encounters a suspicious child, they can take a photo and send it to the concerned authorities. Then the authorities can verify the juvenile through an application developed using the trained model. The face recognition model in our system find a match in the database with the help of face encoding. It is performed by comparing the face encoding of the uploaded image to those of the images in the database. If a match is found, it will be notified to the police station along with the location of where the juvenile is found in order to facilitate their safe return. Our system represents a promising approach to addressing the issue of missing juveniles. It has the potential to greatly improve the speed and efficiency with which they are located and return them to the juvenile home. This paper suggests a framework that would help the police and general society to identify missing juveniles by utilizing a hybrid CNN model. The proposed CNN model have achieved an accuracy of 95%.

Keywords : Juvenile; CNN; Face Recognition; HOG;