Modeling of Artificial Intelligence Enabled Crowd Density Classification for Smart Communities

Smart cities are a contemporary phenomenon to involve information and communication technologies (ICTs) in the advancement of large urban cities. It will be helpful in determining the movement of a city through observing general flow of visitors and traffic jams. Crowd management can be one key aspect of smart cities, assisting in enjoyable and safety experiences for visitors and residents. As crowd density (CD) classification methods encounter difficulties such as inter-scene and intra-scene deviations, non-uniform density, occlusion and convolutional neural network (CNN) methods were valuable. This manuscript designs a wolf pack algorithm with deep learning enabled crowd density classification (WPADL-CDC) model for smart communities. The presented WPADL-CDC technique assists in improving the quality of life in smart community environment. In addition, the presented WPADL-CDC model employs deep convolutional neural network (DCNN) based densely connected network (DenseNet) model for feature extraction purposes. Moreover, the WPA is exploited for the optimal hyper parameter tuning of the DenseNet201 method. Furthermore, fuzzy radial basis neural network (FRBNN) model can be utilized for the identification and classification of CDs in the video surveillance system. For examining the enhanced CD classification outcomes of the presented WPADL-CDC method, a detailed experimental analysis is performed. The experimental values demonstrate the promising performance of the WPADL-CDC model.