Right now one image can be parsed for one table. Ideally one should be able to give one image with all the tables for one room.