Portal de Periódicos da CAPES

Algorithm for child adoption process using artificial intelligence and monitoring system for children

2024; Elsevier BV; Volume: 26; Linguagem: Inglês

10.1016/j.iot.2024.101170

2543-1536

Cosmina-Mihaela Roșca, Adrian Stancu, Andy - Valentin Ariciu,

Child Welfare and Adoption

The child adoption process consists of a complex procedure in any country. The proposed software product focuses on solving particular adoption-related problems that were underscored by some researchers, such as relationship difficulties between the child and adoptive parents due to the type of matching process, lack of accessibility and transparency, high expenses and time-consuming for adoptive parents, and measuring the child's anxiety or stress in the trial adoption. Thus, the first software component computes the compatibility between the child and the parent, the second assesses the parent's readiness for adoption, and the third employs a bracelet to track the child's heart rate and monitor the progress of the trial adoption in real-time. The application "Adopt Platform" was built using C#, the ASP.NET and Entity Framework for Microsoft SQL database communication, and the Azure tool for artificial intelligence functions for English and Romanian languages. The paper incorporates two case studies, employing numerous tests to assess the accuracy of artificial intelligence services in recognizing forms and conducting sentiment analysis through the Azure tool. The results show that the Azure Form Recognizer effectively discerns data aligned with the Adopt Platform's specifications and handles images taken within a 30 cm range. Nevertheless, it imposes constraints: images should stay within a 30° tilt from the horizontal, be within a 30 cm distance, and not contain unrelated text in the backdrop. When evaluating text lengths across English and Romanian contexts, the Azure Sentiment Analysis exhibited an average performance metric of close to 57%.