Video anomaly detection is a complex task that has numerous applications in video surveillance. It involves identifying unusual patterns or events in a video stream that deviate from the expected or typical behavior. This paper introduces a new framework for supervised video anomaly detection using transfer learning from a pre-trained model. The approach presented in this paper utilizes the MobileNetV2 architecture as a feature extractor, which is further fine-tuned using a small set of annotated data. The fine-tuned model is then utilized to classify video frames into normal or anomalous classes. The suggested methodology is evaluated on benchmark datasets and compared with state-of-the-art methods. The experimental results demonstrate the effectiveness and efficiency of the proposed method in detecting anomalies in videos with high accuracy and low computational cost.