Single-cell imaging is a new technology that provides us information such as location, appearance, and biomarkers of the protein levels of each of the sample cells. For analytical use of the samples and comparing them with each other, cells are first quantified. The process involves mapping each cell to a vector, which describes the cell based on the technology used. The indexing methods are then used to summarize the cell information and convert the set of descriptor vectors into a fixed-length vector. Despite the extensive research done on cell indexing, little research has been done on the use of cell location for profile generation, and existing methods can be improved. Our goal in this project is to use spatial information of cells in order to extract more information from the samples and use them in the profile, which ultimately leads to a better understanding of the conditions of the samples than current methods. The varying number of cells and how they are distributed in the images precludes a direct comparison of the samples. Therefore, in this project, using machine learning and algorithmic methods, we will present a method for indexing the mentioned images that will allow direct comparison of them. Finally, the performance of the proposed model will be evaluated using the prediction accuracy of the labels of the images and compared with the existing methods.