Transfer Learning with binary output

i am trying to determine a specific object is present in the image or not. For that i trained googenet and tried with testdata but i dont know how to make a code that point out the object from various images and make bounding box around it. please help me. thanks