Machine learning technology into e-learning platforms means personalization of education with an adaptable learning experience. With all these technologies in relate, there are also powerful technologies targeting the educational system (as well as e-learning systems), Region-Based Convolutional Neural Networks (R-CNNs) are becoming one of the most powerful tools for image and object detection tasks. This current paper intends to explain the use of R-CNNs in e-learning for increasing learners’ engagement and delivering the material and, accordingly, developing various assessment methodologies. With the new capabilities provided through identification and analyzing visual inputs, educators will have the chance to build smart systems such as automated grading of work, instantaneous feedback about visuals that learners created, and even dynamic interactive learning modules adapted to each individual learner's capability. Using region-based convolutional neural network (RCNN) technology, students’ learning styles are fused to dynamically create learning styles on the basis of machine learning technology, which realizes the e-learning in this research. The emotion-cognition relationship is verified, and the RCNN is directly applied for learning style extraction in personalized e-learning research. This approach relies on statistical learning theory and RCNN technology. Incorporation of emotional along with cognitive interactions into tailored e-learning has enabled the development of funds for human-computer interaction with a positive atmosphere. And, finally, the result indicated that an ensemble classifier is the most suitable model to classify coverage using ground data with a high accuracy.