Retinoblastoma (Rb) is a type of eye cancer that starts in the retina, which is the delicate inner lining of the eye. Rb is more common in young children, although it may also attack adults. Leukocoria (LE) is the most common symptom of this cancer, it appears as an abnormal white pupil in images. Rb is mostly diagnosed by performing an ophthalmic exam, but this is not very effective as it leads to late discovery. Early discovery leads to better response to treatment, thus our project sets out to developing an easy to use mobile application to detect Rb in children from images using the MobileNetV2 architecture, in the earliest stage possible . The model was trained using 420 images across the four class labels (healthy, RR, RL, and RB). The model produced an accuracy score of 63%. This study produced good results in spite of the extremely small dataset, however in future we plan to improve the performance of our application and add new features so our application can be used in hospitals everywhere, and by parents anytime.