RETRACTED CHAPTER: Exploiting Deep Learning Techniques for Autistic Face Recognition

Autism face detection is a crucial problem in today’s era. With the advancement in computer vision and pre-trained models it is now easy to utilize less available images and with the help of weights of already trained model in the similar area. This transfer learning approach can help to accurately recognize the face of autistic kids. It will be very useful for early diagnosis and intervention of this neuro-developmental disorder in kids. Our fine-tuned VGG16 achieved 87% avg. accuracy on test dataset. EfficientNetB0 and EfficientNetB7 provides 79.3% avg. and 82.6% accuracy on test dataset. Furthermore, we tested DenseNet121 and SE-ResNet-152 pre-trained models and after performing fine tuning we got 82.67% avg. and 85.3% accuracy on test dataset. Other than that we used MobileNetv1 which achieved 87.0% avg. accuracy on test dataset. Moreover to increase the performance of our model we tried ensemble of two pre-trained model, i.e., Xception and SE-ResNet, and got 87.3% accuracy. We did comparative analysis of various pre-trained models for ASD/Non-ASD face classification.