Newborns kidnapping, switching, disappearance and illegal adoption are becoming problems at hospitals, birthing centers and other places where several births take place simultaneously. Eliminating the occurrence of childhood illnesses that may be prevented by vaccination (e.g., Polio, Tuberculosis, and Tetanus) is also one of the main objectives of most national, international, and non-governmental health organizations. To enhance vaccine coverage, access to healthcare, and receipt of nutritional supplements, an efficient immunization programmer must keep track of toddlers who have been inoculated and those who have gotten the necessary booster injections during the first three years of life. On the other hand, the fingerprint recognition system is effective at authenticating and identifying adults, but when it comes to newborns and solving toddler problems like immunizations against diseases and nutritional care, it encounters difficulties like poor fingerprint image quality and incorrect registration, reducing the recognition accuracy. To solve this problem, we are going to use Pre- trained model that falls under Convolution Neural Networks and we train our dataset on it, the trained model tested have been on data NITG for 154 newborns and toddlers. The experimental results show that the trained model gave better results than before reaching an accuracy of 100% on the training set and 82.47% on validation.

Biometrics, Fingerprint recognition, Newborns, Toddlers, Pre- trained model, Convolution Neural Networks