Missing child recognition system using deep learning

• Main Authors: Geetha, V., Gomathy, C. K., Kasyap, B., Kumar, C. Hemanth
• Format: Conference Proceeding
• Language: English
• Subjects: Annotations; Children; Convolution; Effectiveness; Machine learning; Recognition; Support vector machines
• ISSN: 0094-243X; 1551-7616
• DOI: 10.1063/5.0212567

The widespread public can publish their photographs of questionable children to a everyday portal with annotations and landmarks. The image is mechanically as compared to the repository’s listed photographs of the lacking child. The enter baby picture is classified, and the lacking kids database is searched for the satisfactory suit. so as to do this, a deep gaining knowledge of version is skilled to properly discover the missing toddler the use of the facial photograph uploaded with the aid of the very last phrase public and the lacking child image database that is furnished. For face identity, the Convolution Neural community, a very effective deep studying technology for photograph-based totally programs, is used. With the assist of a pre-educated CNN model VGG-Face deep structure, face descriptors are extracted from the pictures. In evaluation to ordinary deep getting to know applications, our technique merely employs the convolution community as a high level characteristic extractor, leaving the trained SVM classifier to carry out the kid reputation. After deciding on and properly education the only acting CNN model for face popularity, VGG-Face, it will become a totally deep gaining knowledge of version that is resistant to noise, comparison, image pose, the age of the children, and in advance processes to stand recognition-based lacking toddler identification.