Embodiments are disclosed for perspective object snapping. The method may include detecting a first object in a perspective image that includes one or more vanishing points. The method may further include receiving a second object for insertion into the perspective image. The method may further include extracting a plurality of line segments from the first object. The method may further include generating, from the plurality of line segments, one or more snap points. The method may further include generating a perspective bounding box for the second object based on the one or more snap points, the one or more line segments and the one or more vanishing points. The method may further include inserting the second object into the perspective image based on the perspective bounding box.