Aiming at the center isn't a grasp plan
Given an object's shape and pose, a grasp planner has to choose exactly where the gripper's fingers or suction cup should contact the surface, and that choice determines whether the object gets lifted cleanly or slips, rotates, and falls. For a simple symmetric object like a cube, gripping near the center of two opposite faces works fine. But most real objects aren't simple or symmetric: a hammer has its mass concentrated in the head, a stack of books has an uneven center of gravity, and an oddly shaped bracket might have no two parallel flat faces at all. Aiming at the geometric center of an object's silhouette, without accounting for mass distribution and surface geometry, is a common beginner mistake that produces grasps which look fine in a picture but fail the moment the arm accelerates.
This is why grasp planning is treated as its own subfield rather than a trivial last step after perception. The planner has to search over many candidate contact points and orientations, evaluate each one, and pick the best available option, often under a tight time budget because a warehouse pick needs to happen in a fraction of a second.