Mohammed Ghavami

Novel UWB pulse shaping using prolate spheroidal wave functions

In this paper novel prolate spheroidal wave functions are proposed as pulse shapes for use in impulse radio (ultra-wideband) communications. These classes of functions yields orthogonal pulses and have a constant pulse width regardless of the pulse order. This is an important property since it eliminates inter-symbol interference. An M-ary communications system is considered that employs these pulses, and the generation of these pulses using the eigenfunction form of a self-adjoint operator is proposed. It is also shown that these pulses are suitable for use in pulse position modulation (PPM) ultra wideband (UWB) communication systems.

Wavelet vs. Fourier based UWB Systems

In this paper we analyse the application of wavelet transform as an alternative to the conventional Fourier-based multicarrier UWB systems. In Fourier based multicarrier UWB systems a cyclic prefix (CP) with the same length of the channel impulse response must be added to each symbol in order to convert linear convolution into circular convolution. The CP must consist of identical copies of the transmitted data in each symbol and therefore is a waste of bandwidth and resources. In this paper we present a framework for wavelet based multicarrier UWB systems and it is shown that they do not require the cyclic prefix for transmission and, hence throughput increases. Moreover, a closed form formula is derived to represent convolutions counterpart in the wavelet domain. Finally, a performance comparison of both techniques is provided.

Progressive medical image transmission and compression

Digital radiology places very high demands on the networking and digital storage infrastructure of hospitals. In addition to having quite stringent requirements on the quality of the images displayed to the radiologist, much of the technical challenge resides in the necessity of displaying desired images as rapidly as possible. We present an infrastructure for progressive transmission and compression of medical images, which can refine an initial image by increasing the detail information not only in scale-space, but also in coefficient precision. The approach is based on the embedded zerotree wavelet (EZW) algorithm. This algorithm offers a tremendous amount of flexibility in meeting the bandwidth and image quality constraints in a radiological imaging environment. We propose an application of the EZW algorithm in progressive medical image transmission in which it can specify and control both the resolution constraint and rate constraint. The presented method can provide a framework for lossy image compression, with performance far superior to those provided by the standard JPEG algorithm. Also due to the flexibility of the method we will show how any region of interest of an image can be sent progressively.

Performance of space-time block coding and space-time trellis coding for impulse radio

Ultra wideband (UWB) systems have attracted a lot of research interest lately, owing to their appealing features in short-range mobile communications. These features include low power peer-to-peer transmissions, multiple access communications, high data rates, and precise positioning capabilities. Space-time coding (STC) techniques, such as the block coding scheme, or the trellis coding scheme, are known to be simple and practical ways to increase the spectral efficiency in wireless communications. We aim to combine a pulse position modulation (PPM) impulse radio multiple access (IRMA) system with two different space-time coding techniques. Thus, we developed a space-time block code scheme and adapted a space-time trellis code scheme to UWB signalling, relying on the channel state information (CSI) at the receiver side. Then the added diversity produced by these techniques is exploited to enhance the performance of UWB systems. Analysis is conducted in a typical UWB environment, with fractionally-spaced (FS) coherent RAKE detection.

Multi rate/resolution control in progressive medical image transmission for the region of interest (ROI) using EZW

Current progressive image transmission (PIT) systems can not control both bit rate and resolution constraints in images. Due to the partially localised nature of the wavelet transform, the value of any pixel in the image depends on only a small number of wavelet coefficients. Thus, it is possible to specify an arbitrary region of the image and prevent that region from being badly degraded during the compression process. This can be done by representing coefficients corresponding to those regions in the wavelet space more accurately. This is accomplished by multiplying those values by an arbitrary factor before applying the embedded zerotree wavelets (EZW) coding technique. The EZW algorithm is then effectively applied to, wavelet coefficients to refine and encode the most significant ones in the wavelet space. The information of the image is thus transmitted in several successive stages. It is shown that in a PIT system not only any region of interest (ROI) can be sent and received progressively, but also bitrate and resolution constraints can be controlled simultaneously. This option becomes very effective in interactive communication channels where the available bandwidth is at a premium.

UWB multiple-pulse generator and transmitter

In this paper, a multi-pulse generator which generates four different prolate spheroidal wave functions (PSWF), based on a source signal for use in ultra-wideband (UWB) communication systems is proposed, which is then applied to an M-ary communication system. This class of pulse shape yields orthogonal pulses that have a constant pulse width and bandwidth regardless of the pulse order, which is in contrast to the majority of other orthogonal pulse classes. Due to the presented results, it is now possible to build a cheap and easily reproducible UWB pulse generator.

A novel powerloss model for short range UWB transmissions

Narrowband pathloss models are not applicable to the link budget calculation of ultra wideband (UWB) systems. In this paper, a rigorous analysis reveals the powerloss behavior of short-range UWB systems when operating in a cluttered environment. The contribution of this paper is the derivation of a power law model that exhibits a novel breakpoint within the vicinity of the transmitter, which is shown to influence the UWB link budget by up to 5 dB.

IEEE EMBS Asian-Pacific Conference on Biomedical Engineering

Current progressive image transmission (PIT) systems can not control both bit rate and resolution constraints in images. Due to the partially localised nature of the wavelet transform, the value of any pixel in the image depends on only a small number of wavelet coefficients. Thus, it is possible to specify an arbitrary region of the image and prevent that region from being badly degraded during the compression process. This can be done by representing coefficients corresponding to those regions in the wavelet space more accurately. This is accomplished by multiplying those values by an arbitrary factor before applying the embedded zerotree wavelets (EZW) coding technique. The EZW algorithm is then effectively applied to, wavelet coefficients to refine and encode the most significant ones in the wavelet space. The information of the image is thus transmitted in several successive stages. It is shown that in a PIT system not only any region of interest (ROI) can be sent and received progressively, but also bitrate and resolution constraints can be controlled simultaneously. This option becomes very effective in interactive communication channels where the available bandwidth is at a premium.

A new paradigm for telehealth implementation

Recently, efforts have been made to treat patients at home as much as possible. In many cases, the reason for the patient staying in hospital is not that he/she actually needs active medical care. Often, the principal reason for a lengthy stay in hospital is simply continual observation. This paper explains our experience and strategy to support the treatment of patients in their own home through the remote monitoring of physiological signals. The benefits of such remote monitoring are wide-ranging; the patient can continue to live their normal life, their risk of infection is reduced, costs are significantly decreased for the hospital, and clinician time is utilised more effectively.

A matrix representation to increase the resolution of a Region of Interest (ROI) for medical images

Diagnostic quality medical images consume vast amounts of network time, system bandwidth and disk storage in current computer architectures. There are many ways in which the use of system and network resources may be optimised without compromising quality of diagnostic image. One of these is in the choice of image representation, both for storage and transfer. In this paper, we show how a particularly flexible method of image representation can lead to an efficient progressive image transmission method. We present the application of a progressive transmission scheme to medical images, and provide some examples of image refinement in a multi-scale fashion. A matrix approach for the entire multiresolution decomposition of images is also presented. Moreover, we prove that thumbnail images created by a multiscale orthogonal decomposition can be optimally interpolated in a minimum square 2-norm sense, based on a generalised Moore-Penrose inverse operator. These interpolators are the minimum energy interpolators which satisfy the undetermined matrix equation in the multiresolution framework. In addition, we present examples of region of interest (ROI) enhancement in which a user selects a square ROI on a low resolution interpolated image.