Method and system for video processing

A system for video processing, said system comprising:

• a video processing device that comprises circuitry configured to:
• generate a spatial saliency map based on spatial information associated with a current frame of a video stream;
• generate a spatio-temporal saliency map based on at least motion information and spatial distance information associated with said current frame and a previous frame of said video stream;
• assign at least one spatial saliency value to at least one pixel of a plurality of pixels in said current frame based on a difference between an average color value of said plurality of pixels and a color value of said at least one pixel that is more than a threshold value;
• compute at least one combined saliency value of said at least one pixel in said current frame at least based on said at least one spatial saliency value in said current frame;
• and
• generate a combined saliency map based on said at least one combined saliency value and a weighted combination of said generated spatial saliency map and said generated spatio-temporal saliency map.

A non-transitory computer-readable storage medium having stored thereon, computer-executable instructions for causing a computer comprising one or more processors to execute operations, the operations comprising:

• generating, by a video processing device, a spatial saliency map based on spatial information associated with a current frame of a video stream;
• generating, by said video processing device, a spatio-temporal saliency map based on at least motion information and spatial distance information associated with said current frame and a previous frame of said video stream;
• assigning at least one spatial saliency value to at least one pixel of a plurality of pixels in said current frame based on a difference between an average color value of said plurality of pixels and a color value of said at least one pixel that is more than a threshold value;
• computing at least one combined saliency value of said at least one pixel in said current frame at least based on said at least one spatial saliency value in said current frame;
• and
• generating, by said video processing device, a combined saliency map based on said at least one combined saliency value and a weighted combination of said generated spatial saliency map and said generated spatio-temporal saliency map.

A system for video processing, said system comprising:

• a video processing device that comprises circuitry configured to:
• generate a spatial saliency map based on spatial information associated with a current frame of a video stream;
• generate a spatio-temporal saliency map based on at least motion information associated with said current frame and a previous frame of said video stream;
• generate a combined saliency map based on a weighted combination of said generated spatial saliency map and said generated spatio-temporal saliency map;
• and
• determine at least one salient object based on said generated combined saliency map,
• wherein said weighted combination of said generated spatial saliency map and said generated spatio-temporal saliency map is determined based on at least a summation of a first value, a second value, and a third value,
• wherein said first value is a product of a first constant weight and a spatial saliency value of a pixel of said current frame,
• wherein said second value is a product of a second constant weight and a spatio-temporal saliency value of said pixel, and
• wherein said third value is a product of a third constant weight, said spatial saliency value, and said spatio-temporal saliency value.