Tehran University of Medical Sciences Frontiers in Biomedical Technologies 2345-5837 13 1 2026 01 20 199 205 Farah Flayyeh Alkhalid Control and Systems Engineering Department, University of Technology, Baghdad, Iraq 2024 02 19 2024 06 09 Purpose: People with Down Syndrome must be served special because they have an intellectual disability with abnormality in memory and learning, so, creating a model for DS recognition may provide safe services to them, using the transfer learning technique can improve high metrics with a small dataset, depending on previous knowledge, there is no available Down syndrome dataset, one can use to train. Materials and Methods: A new dataset is created by gathering images, two classes (Down=209 images, non-Down=214 images), and then expanding this dataset using Augmentation to be the final dataset 892 images (Down=415images, Non-Down=477 images. Finally, using a suitable training model, in this work, Xception and Resnet models are used, the pretrained models are trained on Imagenet dataset which consists of (1000) classes. Results: By using Xception model and Resnet model, it concluded that when using Resnet model the accuracy = 95.93% and the loss function =0.16, while by using Xception model, the accuracy =96.57% and the loss function =0.12. Conclusion: A transfer learning is used, to overcome the suitability of dataset size and minimize the cost of training, and time processing the accuracy and loss function is good when using Xception model, in addition, the Xception metrics are the best by comparing with the previous studies. https://fbt.tums.ac.ir/index.php/fbt/article/view/943 https://fbt.tums.ac.ir/index.php/fbt/article/download/943/443