Prominent region detection in scenes from sequence of image frames

A method, comprising:

• in an electronic device, comprising an image sensor and circuitry:
• capturing, by the image sensor, a sequence of image frames of a scene;
• receiving, by the circuitry, the sequence of image frames from the image sensor;
• generating, by the circuitry, an optical flow map for a current frame of the sequence of image frames, wherein
• the optical flow map comprises a plurality of regions, and
• the plurality of regions corresponds to a plurality of moving objects within a field-of-view of the image sensor;
• determining, by the circuitry, a plurality of likelihood-of-prominence values for the plurality of regions based on the generated optical flow map and a device motion vector, wherein the device motion vector corresponds to a direction of motion of the electronic device;
• detecting, by the circuitry, at least one region from the plurality of regions as a region-of-prominence based on the likelihood-of-prominence values, wherein
• a direction of motion of the at least one region is same as the direction of motion of the electronic device, and
• the at least one region maps to a desired object of the plurality of moving objects;
• and
• tracking, by the circuitry, the desired object in the sequence of image frames based on the detection of the at least one region of the current frame as the region-of-prominence.