Image processing method and apparatus

A method of processing image data representing an image, the method comprising steps stored on a tangible computer readable medium and performed by a processor, the steps comprising:

• generating saliency data by determining a series of features of the image and using the determined features to generate a probability measure for each point of the image representative of a location of a subject of the image;
• convolving the saliency data with a set of functions of respectively different spatial scales in order to generate response data representing a saliency density for the image at respective ones of the spatial scales;
• determining from the response data spatial maxima in the saliency density across the respective scales for respective ones of the determined features;
• and
• selecting, using the spatial and scale location of a determined maxima, image data representing a salient image portion at that scale.

A method of processing image data representing an image captured using an image capture device, the method comprising steps stored on a tangible computer readable medium and performed by a processor, the steps comprising:

• processing the image data in order to generate saliency data representing salient portions of the image;
• processing the saliency data at a plurality of spatial scales using respective ones of a plurality of weighting functions, the processing operable to generate response data representative of a measure of the relative saliency of portions of the image at respective ones of the spatial scales;
• using the response data in order to determine a preferred scale for the salient portions of the image by determining maxima in the response data for respective ones of the salient portions, and generating region data representing respective ones of the salient portions at their preferred scales;
• using the region data to generate a visualization path which substantially traverses respective ones of the salient portions, wherein the path is generated on the basis of the scale of the salient portions, and the relative distance between salient portions.

Apparatus comprising a processor operable to process image data representing an image, in accordance with computer executable program codes stored on a tangible computer readable medium, the processor operable to:

• generate saliency data for the image by determining a series of features of the image and using the determined features to generate a probability measure for each point of the image representative of a location of a subject of the image;
• process the saliency data using respective ones of weighting functions of a plurality of spatial scales in order to determine the positions of regions of interest of the image at respective ones of the scales;
• generate response data for each scale representing the relative strength of response of the saliency data at the positions of the determined regions of interest to the function at that scale;
• for each determined region of interest, process the response data to determine a maximum response over all scales for that region;
• and
• use the position in the image corresponding to the determined maximum and the scale at which the maximum occurs to generate output image data representing a salient portion of the image at that scale which corresponds to the region of interest in question.

A computer program product for use with a computer, said computer program product comprising a tangible computer readable medium having computer executable program code embodied thereon, wherein said product is operable, in association with said computer, to:

• generate saliency data using image data representing a captured image by determining a series of features of the image and use the determined features to generate a probability;
• measure for each point of the image representative of a location of a subject of the image;
• process the saliency data using respective ones of weighting functions of a plurality of spatial scales in order to determine the positions of regions of interest of the image at respective ones of the scales;
• generate response data for each scale representing the relative strength of response of the saliency data at the positions of the determined regions of interest to the function at that scale;
• for each determined region of interest, process the response data to determine a maximum response over all scales for that region;
• and
• use the position in the image corresponding to the determined maximum and the scale at which the maximum occurs to generate output data representing a salient portion of the image at that scale which corresponds to the region of interest in question.