Content based image selection for automatic photo album generation
AssigneeXerox Corporation
InventorClaudio Cifarelli
Life of this patent
Estimating the term…
Abstract
A method is employed to present image from an event. A plurality of images from the event are received and one or more clusters of images are created, wherein each cluster of images has a similarity greater than a predetermined threshold. A density value of the data distribution within each cluster is estimated and at least one local maximum associated with each cluster is identified via a density function. At least one image from each cluster is selected, wherein each image is a candidate for the presentation. A layout is created to present the images selected.
Claims19
A method for generating an album describing an event, the method comprising:
- receiving a plurality of images in an image data set;
- evaluating a similarity of images based on one or more metrics associated with each image;
- creating a cluster of images having a similarity score greater than a predetermined threshold;
- identifying local maximum within the data set via a density function;
- using each local maximum to identify an element representing a peak moment within a given event;
- selecting one image within each closed neighborhood of images representing the local maximum by considering one or more disparate features, the disparate features including at least one of image quality, time of capture, user-rating, and ranking;
- selecting a template for laying out the selected images, the selecting related to criteria associated with each image;
- assigning the each image to the template based on a time associated with the each image;
- and,
- presenting the selected images in the layout.
A method to provide a representative compilation of images from an event, comprising:
- receiving a plurality of images in an image data set;
- evaluating a similarity of images based on one or more metrics associated with each image;
- determining a density of each image within the data set using the equation, dens(i)=Σh=1Rn(i,h)/V(h), wherein n(i) is a number of points that fall in a hypersphere with a radius of rh with a volume V(h);
- identifying a local maximum for each hypersphere within the data set using the density;
- using each local maximum to identify an element within the images representing a peak moment within a given event;
- selecting one image within each closed neighborhood of images representing the local maximum by considering one or more disparate features;
- creating a layout for presenting the selected images;
- and,
- presenting the selected images in the layout.
A system that identifies images for presentation within a compilation, comprising:
- an image data set that contains a plurality of images;
- a similarity evaluator that receives the image data set, determines a similarity score for each image relative to every other image within the data set and assigns each image to a cluster based on the similarity score;
- a density estimator that identifies local maximum within the data set via a density function and uses each local maximum to identify an element representing a peak moment within a given event;
- an image selector that selects one image within each closed neighborhood of images representing the each local maximum by applying at least one of an image selection algorithm for eliminating images not captured within one of a predefined window and a ranking algorithm for ranking remaining images;
- and,
- a layout creater for placing the selected images in a template.
Dependencies are inferred from the claim text.