Evaluating Automated Face Recognition Tools on a Head-Mounted Camera Dataset of Young Children with and without Down Syndrome
Keywords:
Head-mounted camera, face recognition tools, Down syndrome, visual experiences, automated analysis, computer vision, neurodevelopmental conditions, neurodiversity.Abstract
Young children actively shape their learning through visual exploration, butneurodevelopmental conditions like Down syndrome may influence this process. This study assesses automated face recognition tools using a novel dataset from head-mounted cameras worn by young typically developing children and those with Down syndrome. Recorded in home environments, the dataset includes challenges such as varied angles and motion blur. Multiple recognition tools were tested on thousands of images extracted from video footage. Results show that certain tools performed well, suggesting their potential for automating data analysis in neurodiverse populations. These findings support the development of scalable tools for early intervention. Future efforts will refine these tools with a larger dataset.
