Early detection of ASD Traits in Children using CNN

Autism is neurological disorder in which person is affected with communication and interaction abilities. Lacks of social interaction, repetitive behavior, and stable interest are indication of the autistic child. It essential to identify the autism at very is early stage. CNN plays vital role in health care which requires a process that reduces cost and time. The key objective of proposed paper is to implement convolution neural network algorithms and classify autistic and non-autistic child..In this study, CNN is applied for classification of autistic and non-autistic child. The images of children of age 4 to 11 years were used. About 400 images extracted from pre-defined datasets and were used to train the CNN algorithm using the Google colab framework via Python and Open CV libraries. Using cross validation techniques, The CNN was evaluated. In this sense, our proposed model has achieved a high accuracy rate and robustness for prediction of autistic and non-autistic child. Additionally, the proposed algorithm attains a quick response time. Therefore, we could significantly diminish the time of diagnosis by applying the proposed method and facilitate the diagnosis of ASD in lower cost.