Smart and technical identification in crowed places more advanced development and challenging part of deep neural networks. Deep learning approaches are currently being employed to transform traditional human behavior recognition in video surveillance into intelligent techniques. To adopt safety measures in public gatherings, this paradigm shift offers numerous sophisticated features. Proactive surveillance, identification, and supervision of diverse crowd gatherings can improve many crowd-management-related operations in terms of usefulness, capacity, predictability, and safety. Convolutional neural networks, in spite of a number of issues such as occlusion, clutter, uneven item distribution, and non-uniform object scale, are a promising method for recognizing human actions in huge data, such as CCTV. Hence, this study aims to propose a CNN-based Abnormal Classifier model for identifying the abnormal behavior of humans in public gatherings such as Airports, Malls, and Theaters. The model of our proposed on the surveillance dataset and the data analyzed to increase its efficiency without preprocessing approaches for the particular field. The proposed model has the ability to classify and obtain the final behavior classification results. The experimental depicts that the proposed method marginally outperforms the existing methods.