In this chapter, we propose a novel framework, FaceComm for identifying communities from group photographs using face recognition system (FRS), clustering, and community detection algorithms. We evaluated the robustness of the pre-trained facial feature (model) by combining it with Infomap (face clustering) and testing across various test datasets, i.e., MS1M, VGGFace2, and CASIA. Our approach demonstrated average accuracy gains of 4.46% and 4.49% on the pairwise and BCubed F-scores, respectively, across all test datasets. These results surpassed those of other methods, indicating a notable improvement in clustering performance. Since our proposed work is novel, no standard dataset was available to the best of our knowledge. Therefore, to showcase the effectiveness of the proposed framework in a variety of real-world scenarios, we created a custom dataset by systematically collecting 342 images from the Internet, featuring multiple subjects per image in various settings. The images represent a diverse range of backgrounds, including politicians, business professionals, celebrities, and sports personalities. The FaceComm begins by FRS extracting facial features, which are subsequently clustered to identify unique faces across photographs. A graph G = (V, E) is then constructed, where the vertices V represent unique faces, and weighted edges E are assigned between faces that appear together in a single photograph. Individuals (unique faces) who appear frequently in photographs will have higher edge weights, which enhances modularity. Finally, a community detection algorithm is applied to this graph to identify distinct face communities. By visually plotting the identified communities, we can observe the distinct social heterogeneity among unique individuals across categories. The FaceComm effectively uncovers social structures within group photographs providing valuable insights that can be highly beneficial in fields such as security, surveillance, and law enforcement.