<p>The assessment is one of the most significant factors in offline classroom teaching set up. The overall assessment of a classroom session is a vital component to analyse the effectiveness of the classroom teaching. Moreover, interactiveness of a classroom session which takes place between instructor and listeners is a quality measure that can indicate the effectiveness of the outcome for the class. The traditional mechanism to measure the degree of interactiveness is measured via various studies that mainly rely on human perception that may suffer from various issues such as human bias, inaccuracy, and efficiency. Besides, an expert evaluator is required to carefully examine the overall efficacy of classroom teaching learning process. The emergence of artificial intelligence (AI) inspired methods augmented with computer vision paradigm facilitate to design more efficient tools in the field of contemporary assessment tool. In this research, we develop an automatic AI-based approach to measure the effectiveness of a classroom session as either interactive or non-interactive. We use the notion of video analyses along with pre-trained VGG-19 to assess various activities involving presenter and learner in the classroom session. A model is trained on various scenarios of classroom interactive and non-interactive activities. Our model accepts a classroom video session of 50–60&#xa0;min and categories it as interactive or non-interactive class. The proposed model is trained on a novel self- created “CUJ-CS_v.01” dataset and validated on several real-time classroom sessions. The presented method demonstrates its outstanding performance as compared to traditional methods with an accuracy of 99.88% in known scenarios and 98.98% in unknown scenarios. Moreover, our approach can intelligently classify a given video lecture as interactive or non-interactive within few seconds.</p>