Apparatus and methods for saliency detection based on color occurrence analysis

A non-transitory computer-readable storage medium comprising computer readable instructions embodied thereon, that when executed by a processor, cause the processor to:

• receive at least one image from at least one sensor, the at least one image comprising a plurality of pixels;
• partition a first image, of the at least one image, into at least one partition image comprising a distribution of the pixels based on saliency parameters of each pixel;
• use the at least one partition image to determine a saliency map based on areas of high saliency found within the at least one partition image, wherein the saliency map comprises areas of high saliency detected within at least one partition image;
• and
• determine an object mask from the saliency map, wherein the object mask is determined based on at least one area of high saliency within the saliency map exceeding a threshold, indicative of an object within the first image.

A robotic system comprising:

• at least one sensor configured to receive image data of a surrounding environment;
• at least one operative unit configured to receive user input;
• at least one processor configured to execute computer readable instructions to:
• receive at least one image from the at least one sensor, the at least one image comprising a plurality of pixels;
• partition a first image, of the at least one image, into at least one partition image comprising a distribution of the plurality of pixels based on saliency parameters of the plurality of pixels;
• use the at least one partition image to determine a saliency map based on areas of high saliency found within the at least one partition image, wherein the saliency map comprises areas of high saliency detected within the first image;
• and
• determine an object mask from the saliency map, wherein the object mask is determined based on at least one area of high saliency within the saliency map exceeding a threshold, indicative of an object within the first image.

A method for a robotic apparatus to detect objects comprising:

• receiving at least one image from at least one sensor, the at least one image comprising a plurality of pixels;
• partitioning a first image, of the at least one image, into at least one partition image comprising a distribution of the pixels based on saliency parameters of each pixel;
• using the at least one partition image to determine a saliency map based on areas of high saliency found within the at least one partition image, wherein the saliency map comprises areas of high saliency detected within the first image;
• and
• determining an object mask from the saliency map, wherein the object mask is determined based on at least one region of pixels of high saliency within the saliency map exceeding a positive detection parameter threshold, indicative of an object within the first image.