|
|
|
William T. Plummer 01/20/2008 These comments address the several technologies originated by RGB Light ltd for "enhancement and compression" of still and video digital images. Within the material presented I have identified a means of lossless compression, a means of lossy compression, and at least three technologies that may be useful in enhancing digital images. Lossless Compression. The lossless compression method, called the “New Byte Representation”, uses a change in mathematical representation to improve on conventional methods of lossless compression. Conventionally, lossless compression identifies sequences of unchanging values in the color magnitudes within a stream of pixels, and replaces the repeating individual magnitudes with a simple statement of the number of subsequent pixels that repeat that magnitude. What Massimo Ballerini has done is to increase the amount of redundancy available for this type of compression by partitioning each color magnitude into at least one “coarse” part and at least one “fine” part, to be dealt with separately. In most photographs there will be longer sequences of unchanging coarse parts, and this increased redundancy has been exploited to provide a greater factor of lossless compression. Lossy Compression. The lossy compression method, called the “Dark Method”, offers a larger compression ratio than the lossless method, by sacrificing some of the information in an original digital photograph or video stream. Unlike lossless compression, in which every compression method provides a route to restore exactly the original picture when compression is reversed, different methods of lossy compression will restore an expanded picture with different kinds of information loss, and with different resulting artifacts that may compromise the visual quality. Massimo Ballerini’s Dark Method sacrifices the smaller numerical gradations in the color values at each pixel, while maintaining full spatial resolution in the picture after expansion. Although a similar result may be achieved by retaining fewer distinct color levels in the initial acquisition of the picture, the dark method begins with the conventional 256 or more levels for each color, and therefore provides the opportunity for nonlinear spacing of the reduced number of levels retained in the compression step, a way to provide higher picture quality at any selected amount of compression. Because the Dark Method does not directly exploit the eye’s limited ability to resolve fine spatial detail in chrominance, for example, I anticipate that the Dark Method would not normally be used alone, but would be combined with one or another of the more conventional means of lossy compression for a better joint result. Enhancement. Bidimensional Color Space, Tetrahedric Model, and Quadratic Color. These three technical proposals offer means of enhancing images, rather than compressing them. There advantage over previous art appears to be in providing easier and more efficient processes for enhancing the appearance of the images. In some cases a given kind of enhancement can be achieved with fewer individual manipulations than would be needed in the conventional use of image enhancement software. Market. Potential for market applications I believe that these technologies offer advantages over methods now conventionally used, and deserve to be protected individually by patent. The lossless compression will be easiest to demonstrate to potential users because all lossless methods will return the original image when reversed. The methods will be differentiated by two metrics: computer time required to produce and reverse the compression, and degree of compression achieved. The degree of compression achieved is of the greatest economic importance because transmission bandwidth is limited in most cases, and is expensive. Greater compression is reflected immediately in reduced transmission cost for an image or sequence of images. Lossy compression by the dark method is more difficult to compare with other means of lossy compression simply because the different means result in different kinds of loss or artifact in the expanded final image. But comparison can still be made through a joint evaluation of degree of compression together with a quality metric, such as PSNR. On this basis the dark method can be compared with alternative lossy compression procedures, alone or in combination with other means. The commercial value will again be principally the saving in transmission bandwidth, and lossy compression will require much less bandwidth than lossless compression. I believe that the three picture enhancement technologies may be most attractive if packaged as “plug-ins” with other software provided for manipulating and enhancing digital images. Market Opportunities Any new compression technology must provide for both compression and expansion back to a final image. When a new technology is introduced, it is most easily introduced in a situation where one individual or one machine will be responsible for or will carry out both processes. For lossless compression one such application is the compression, transmission, and restoration of medical images, such as digital x-rays. Such images are now transmitted around the world in large numbers for reading and discussion by radiologists on other continents to take advantage of staggered time zones. Example: the “Nighthawks” who read x-rays in off hours. In this application the image quality is unusually demanding and the cost of transmission may be high. But the sender and the receiver are in easy communication with each other, and may quickly try an innovative technology without first achieving agreement on a new general standard. I would suggest seeking out and working with individuals engaged in this work to compare Ballerini’s new lossless compression with other means that have been tried. Bandwidth advantages will be immediately visible, and this demanding use should be a source of valuable testimonials. A comparable opportunity for lossy compression exists with the TiVo hard-disk video recorder. Lossy compression and decompression are carried out within a single box, so no wide-scale agreement or adoption of a new standard would be needed. Lossy compression is currently offered with a range of final quality loss, to be traded off against the capacity of the storage device. A more ambitious use of either lossless or lossy compression will require implementation of a new standard for the purpose. One business model is the .pdf format used by Adobe Acrobat; decoding software is distributed free for easy downloading, and Adobe sells encoding software. Because Ballerini’s lossless and lossy compression technologies are most effective for image cases different from those most commonly used by Acrobat, it is possible that Adobe (or an Adobe competitor) might find these new technologies attractive in an extension for that existing business. Adobe Photoshop and its related software products may also be a route for getting Ballerini’s image enhancement software into broad application with only minor investment. The Next Steps The most immediate need will be to describe the new technologies more completely in a way that will facilitate a more comprehensive patent search and final applications. If there are related technologies in use, we want to find them quickly. At the same time, these new compression technologies must be benchmarked quantitatively against other “best practice” technologies, to demonstrate advance over existing art both for proof of novelty and for commercial market purposes. Bill Plummer at www.wtpoptics.com |