Technical Report CS-2011-10

Title: Adaptive Image Compression Using Sparse Dictionaries
Authors: Inbal Horev, Ori Bryt and Ron Rubinstein
Abstract: Transform-based coding is a widely used image compression technique, where entropy reduction is achieved by decomposing the image over a dictionary known to provide compaction. Existing algorithms, such as JPEG and JPEG2000, utilize fixed dictionaries which are shared by the encoder and decoder. Recently, algorithms using content-adaptive dictionaries have been proposed. These algorithms employ more specialized dictionaries which are optimized for specific classes of images, and in this way are able to improve compression rates for these classes. Such approaches, however, lose generality as they require sharing a specialized dictionary in advance between the encoder and decoder, for every type of input image.

Utilizing image-adaptive dictionaries has the potential of restoring generality and improving compression rates by encoding any given input image over a dictionary specifically adapted to it. However, this has so far been avoided as it requires transmitting the dictionary along with the compressed data.

In this work we use the sparse dictionary structure to implement generic image compression. This dictionary structure has a compact representation, and thus can be transmitted with relatively low overhead. We employ this structure in a compression scheme which adaptively trains the dictionary for the input image. Our results show that although this method involves transmitting the dictionary, it remains competitive with fixed-dictionary schemes such as JPEG and JPEG2000.

CopyrightThe above paper is copyright by the Technion, Author(s), or others. Please contact the author(s) for more information

Remark: Any link to this technical report should be to this page (, rather than to the URL of the PDF files directly. The latter URLs may change without notice.

To the list of the CS technical reports of 2011
To the main CS technical reports page

Computer science department, Technion