In this project, Semantic segmentation model of surgical instruments in robotic-assisted surgery has been made. In other words, the goal of your work is to identify the instruments in a surgical scene and correctly labeling each instrument pixel with its class. The segmentation of surgical instruments or their type is frequently used as an intermediate task for the development of computer-assisted surgery systems such as instrument tracking, pose estimation, and surgical phase estimation, which in turn have applications ranging from operating room optimization to personalization of procedures, and particularly, in pre-operative planning. Hence, developing reliable methods for the semantic segmentation of surgical instruments can advance multiple fields of research. In this project, Dataset used is from Robotic Scene Segmentation Sub-Challenge of Endovis 2018.