Systems and methods for optimizing a scene

A computer-implemented method comprising:

• defining at least one visual goal for a scene, the scene comprised of a plurality of objects, the visual goal specifying at least one object in the scene which is desired to be not predicted to be attended to by a visual attention model;
• receiving input defining allowable changes to the scene, which define how at least some of the plurality of objects may be changed in order to achieve the goal;
• assigning a plurality of costs to at least some of the allowable changes, each of the plurality of costs associated with one of the allowable changes;
• automatically modifying, by a processor, some of the plurality of objects consistent with the defined allowable changes so as to produce modified scenes;
• evaluating the modified scenes with a visual attention model;
• generating, by the processor, a function value by using an utility function with the plurality of costs and the evaluation;
• and,
• determining, by the processor, based on the evaluation and the function value, which of the modified scenes achieves the at least one visual goal for the scene and has the modification cost being lower than a predetermined threshold.

A computer system comprising:

• a processor and memory;
• a visual attention module operative to evaluate a scene and predict objects within the scene that will tend to attract visual attention;
• and
• a scene optimization module operative to at least:
• receive input defining at least one visual goal for a scene, the scene comprised of a plurality of objects, the objects having properties;
• receive input defining allowable changes, which define how properties can be changed;
• associate allowable changes with a plurality of costs, each of the plurality of costs associated with one of the allowable changes;
• generate modified scenes by modifying the properties of objects consistent with the allowable changes;
• determine the costs associated at least some of the modified scenes based on the plurality of costs;
• generate a function value by using an utility function with the plurality of costs and the visual attention module prediction;
• and
• interact with the visual attention module to determine which modified scenes meet the visual goal.