Development of Facial Recognition-Based Toddler's Emotion Prediction System
Saved in:
| Title: | Development of Facial Recognition-Based Toddler's Emotion Prediction System |
|---|---|
| Language: | English |
| Authors: | G. K. Afolabi-Yusuf, B. A. Bashiru, F. A. Odutayo |
| Source: | Southeast Asia Early Childhood. 2025 14(2):103-110. |
| Availability: | National Child Development Research Centre. Universiti Pendidikan Sultan Idris, 35900 Tanjong Malim, Perak, Malaysia. e-mail: ncdrc@upsi.edu.my; Web site: https://ejournal.upsi.edu.my/index.php/SAECJ/Home |
| Peer Reviewed: | Y |
| Page Count: | 8 |
| Publication Date: | 2025 |
| Document Type: | Journal Articles Reports - Research |
| Descriptors: | Toddlers, Recognition (Psychology), Human Body, Emotional Response, Visual Stimuli, Algorithms, Psychological Patterns, Predictor Variables, Accuracy, Artificial Intelligence, Child Behavior |
| ISSN: | 2821-3149 |
| Abstract: | Young children express their feelings through facial or verbal expressions that differ from person to person and are shaped by the environments where they live. Neglecting to understand and estimate how toddler emotions change can lead to delayed intervention timing, resulting in harm to their mental and social development processes. The study initiated the development of a system based on facial recognition processes to forecast toddler emotional responses. The random forest algorithm was used to build the system model, which received training from a dataset comprising 2,168 pictures showing both facial expressions of happiness and sadness. Mediapipe, a machine learning algorithm, was used for feature extraction. The model was then integrated into a user-friendly interface designed for ease of use. This interface captures a toddler's facial image and classifies their emotion as either happy or sad. In conclusion, the developed model demonstrated strong performance, achieving an accuracy of 84%. By providing real-time emotion predictions, the system can assist parents and caregivers in responding appropriately to a toddler's emotional state. |
| Abstractor: | As Provided |
| Entry Date: | 2026 |
| Accession Number: | EJ1498224 |
| Database: | ERIC |
| Abstract: | Young children express their feelings through facial or verbal expressions that differ from person to person and are shaped by the environments where they live. Neglecting to understand and estimate how toddler emotions change can lead to delayed intervention timing, resulting in harm to their mental and social development processes. The study initiated the development of a system based on facial recognition processes to forecast toddler emotional responses. The random forest algorithm was used to build the system model, which received training from a dataset comprising 2,168 pictures showing both facial expressions of happiness and sadness. Mediapipe, a machine learning algorithm, was used for feature extraction. The model was then integrated into a user-friendly interface designed for ease of use. This interface captures a toddler's facial image and classifies their emotion as either happy or sad. In conclusion, the developed model demonstrated strong performance, achieving an accuracy of 84%. By providing real-time emotion predictions, the system can assist parents and caregivers in responding appropriately to a toddler's emotional state. |
|---|---|
| ISSN: | 2821-3149 |
