Lymphocytes have been found to be important biomarkers in multiple tumors. However, automating lymphocyte counting on tissue images is not a trivial task. These tissue images are massive in size, requires a certain set of rules to be followed when being analysed and moreover, the counts produced by different annotators for the same tissue often are not tallying. Given the prognostic relevance of lymphocyte counts in treating cancer patients, having a unified framework which can overcome these limitations is highly appreciated by the medical staff. In my research I am trying to address these challenges and build a reliable and robust model to assist doctors in their clinical activities.