Multiple image storage compression tree

Abstract

A method, system, and computer program product for compressing an image using similar images includes: receiving a first image; storing the first image on a storage server; comparing the first image to one or more stored intra-frames (I-Frames) to determine a similar I-Frame from the one or more stored I-Frames; in response to determining the similar I-Frame, determining that one or more stored predicted frames (P-Frames) reference the similar I-Frame; comparing the first image to the one or more stored P-Frames to determine a similar P-Frame; determining whether the first image meets a P-Frame threshold level for the similar P-Frame; in response to determining that the first image meets the P-Frame threshold level, generating a first P-Frame for the first image using data from the similar P-Frame and data from the similar I-Frame to compress storage space used by the first image on the storage server.

1 . A computer-implemented method comprising:
 receiving an image, wherein the image is an uncompressed image;   storing the image on a storage server, the storage server storing a plurality of images;   comparing the image to one or more stored intra-frames (I-Frames), each of the one or more stored I-Frames corresponding to an image from the plurality of images, to determine a similar I-Frame from the one or more stored I-Frames, wherein comparing the image to the one or more stored I-Frames comprises:
 determining similarity values for each of the one or more stored I-Frames, 
 comparing each similarity value from the similarity values to a similarity threshold, wherein the similarity threshold is a number indicating a threshold amount of variance between the image and the one or more stored I-Frames, and wherein the similarity threshold is specified by an owner of the storage server, and 
 determining a similarity value from the similarity values that is greater than the similarity threshold; 
   in response to determining the similar I-Frame, determining that one or more stored predicted frames (P-Frames) reference the similar I-Frame, wherein determining that the one or more stored P-frames reference the similar I-Frame includes determining that the stored P-Frame was compressed using data from the similar I-Frame;   comparing the image to the one or more stored P-Frames to determine a similar P-Frame, wherein the similar P-Frame is a stored P-Frame, from the one or more stored P-Frames, with a similarity value greater than the similarity threshold;   determining whether the image meets a P-Frame threshold level for the similar P-Frame, wherein determining whether the image meets the P-Frame threshold level comprises:
 determining the similarity value of the similar P-Frame, and 
 comparing the similarity value of the similar P-Frame with the P-Frame threshold level; 
   in response to determining that the image meets the P-Frame threshold level, determining whether the image meets a bidirectional predicted frame (B-Frame) threshold level for the similar P-Frame, wherein the B-Frame threshold level indicates a threshold amount of similarity between the image and the similar P-Frame, wherein determining whether the image meets the B-Frame threshold level includes comparing the similarity value of the similar P-Frame with the B-Frame threshold level; and   in response to determining that the image meets the B-Frame threshold level, compressing the image into a B-Frame using data from the similar P-Frame and the similar I-Frame, wherein the image is compressed into the B-Frame after the image is stored on the storage server for a predetermined time threshold.