نوع مقاله : مقاله کامل پژوهشی
نویسندگان
1 کارشناسی ارشد، گروه مهندسی کامپیوتر، موسسهی آموزش عالی آیندگان، تنکابن، ایران
2 استادیار، گروه انفورماتیک سلامت، پژوهشکدهی تروما، دانشگاه علوم پزشکی گیلان، رشت، ایران
3 دانشیار، گروه مهندسی کامپیوتر، موسسهی آموزش عالی آیندگان، تنکابن، ایران
4 استادیار، گروه انفورماتیک پزشکی، دانشکدهی پیراپزشکی، دانشگاه علوم پزشکی شهید بهشتی، تهران، ایران
کلیدواژهها
موضوعات
عنوان مقاله English
نویسندگان English
IQ test is a common solution to detect and measure people's intelligence, but studies show that the brain activity of intelligent people is significantly different compared to people with normal and low IQ, and this difference can be detected from advanced magnetic resonance imaging or MRI. In this regard, the aim of this paper is to present a model based on artificial intelligence to estimate people's IQ from brain MRI images. To achieve this goal, the combination of Convolutional Neural Network (CNN) and eXtreme Gradient Boosting (XGBoost) is used in this paper. Convolutional neural networks excel in extracting features from images and make them effective for processing visual stimuli related to IQ tests. On the other hand, the XGBoost algorithm is a powerful ensemble algorithm that can effectively combine the predictions of multiple models to improve overall accuracy. In the proposed model, ResNet-50 and VGG16 were used as a feature extractor and the XGBoost algorithm was used as an identifier at the top level of the network to generate results. The proposed model was tested on the Autism Brain Imaging Information Exchange (ABIDE) brain image dataset. Based on the results of the experiments, the use of the pre-trained convolutional neural network VGG16 showed a better performance than the pre-trained convolutional neural network ResNet-50 in the feature extraction section, and its combination with XGBoost algorithm obtained an accuracy of 83% achieved against 60% accuracy of ResNet-50.
کلیدواژهها English