Highly Nonlinear Approximations for Sparse Signal Representation

Logo EPSRC
next next up previous
Next: Image Approximation Up: Examples Previous: Examples

A competitive scheme for storing sparse representation of X-Ray medical images

Sparse image representation refers to techniques for data reduction. Rather than representing the image informational content by the intensity of the pixels, the image is transformed with the aim of reducing the number of data points for reproducing the equivalent information. In a previous work, we have demonstrated that sparse representation, obtained by a large dictionary and greedy algorithms, renders high quality approximation of X-Ray images with high level of sparsity, in comparison with traditional transforms. Here we present a simple scheme which allows to store the sparse approximation of an X-Ray medical image in a file of competitive size with respect to the most commonly used formats, namely JPEG and JPEG2000 (JPEG2)

"A competitive scheme for storing sparse representation of X-Ray medical images''
by Laura Rebollo-Neira

The routines for implementing the methods in the paper and reproducing the results in Table 1 are available here.