Down syndrome is a genetic disorder that affects 1 in every 1000 babies born worldwide. The cases of Down syndrome have increased in the past decade. It has been observed that humans with Down syndrome generally tend to have distinct facial features. This paper proposes a model to identify people suffering from Down syndrome based on their facial features. Deep representation from different parts of the face is extracted and combined with the aid of Deep Convolutional Neural Networks. The combined representations are then classified using a Random Forest-based pipeline. The model was tested on a dataset of over 800 individuals suffering from Down syndrome and was able to achieve a recognition rate of 98.47%.
CITATION STYLE
Mittal, A., Gaur, H., & Mishra, M. (2020). Detection of Down Syndrome Using Deep Facial Recognition. In Advances in Intelligent Systems and Computing (Vol. 1022 AISC, pp. 119–130). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-981-32-9088-4_11
Mendeley helps you to discover research relevant for your work.