Abstract
Deep learning, or artificial neural networks, has been used for years to solve a range of problems. The computation burden increased with the intricacy of challenges and models as well as with the volume of data. This research provides a quick overview of the process of moving beyond artificial neural networks to deep learning, theories, and applications. There's also information regarding used libraries, applications, or electronics. Specifically covered the canonical strategies for crowd analysis. Throughout the study on deep learning approaches for a crowd, analysis is included in the literature as well as an overview of datasets. Studies carried out more recently have also been reviewed and compared. As a result, crowd analysis is an area of inquiry that blends academic theory with real-world application. As a corollary, deep learning is utilized to achieve successful outcomes in the academic and practical realms of crowd analysis.