![]() | Only 14 pages are availabe for public view |
Abstract Image compression lS an important tool to store and transmit visual information that used in several applications such as: satellite, remote sensing, multimedia communications, television broadcasting, internet, etc... . The necessity of compression process is because of the huge amount of transferred data in most of the applications, which exceeds the capability of today’s hardware. Compression of an image refers to a process in which the amount of data used to represent an image is reduced to meet a bit rate requirement (below or at most equal the maximum available bit rate), while the quality of the reconstructed image satisfies the requirements for a certain application and the complexity of the computation involved is affordable for the application. The concept of Progressive Image Transmission (PIT) is of particular importance in browsing large image files. Progressive transmission of an image permits the initial reconstruction of an approximation followed by. a gradual improvement of quality in the image reconstruction. In order to send image data progressively, the data should be organized in pyramidal form according to the order of its importance, from the global characteristics of an image to the local etails. For building this pyramidal organization of data the spatial encoding, or yramidal encoding, is used. Hence, the pyramidal encoding method generates a et of image frames at different resolutions; the image is successively reduced in patial resolution and size by subsampling or averaging. Approximation of an age can be obtained using a single frame or a combination of frames of the age, therefore sending a set of image frames in pyramid form from bottom to p naturally constitutes a progressive transmission. n this thesis, the development of new image compression methods based on erging different Neural Networks (NN’s) and Inverse Difference Pyramidal |