This research proposes a pupil tracking method, which can determine the exact positions of the eyes from live streaming. The introduced method is immune to pose changes, various facial expressions and illumination variations. Initially, it utilizes video frame enhancement, Gabor transformation and its sparseness to extract eye windows. Recent techniques were evaluated against the same comparison data set to check accuracy of the results. The experimental findings demonstrate the consistent robustness and efficiency of the proposed method. This research enables this feature to be integrated with website marketing analysis, smart home control, car accident protection and many other interesting fields.