P2-0065 — Final report
1.
A comparison between SAR data compression in Cartesian and polar coordinates.

A wavelet-based lossy compression method of single-look complex SAR data is presented together with a comparison between compressions in Cartesian and polar coordinates. We shows that the compression of complex SAR data in polar coordinates has smaller amplitude and phase errors than in Cartesian coordinates. The compression in polar coordinates requires bit allocation processing between amplitude and phase. This is achieved by phase scaling before compression.

COBISS.SI-ID: 8687382
2.
Low bit rate wavelet video codec using edge based motion estimation and context-based coding.

We develop a wavelet codec simple 16-bit fixed-point digital DSP. The target application are grabing images from a camera, coding and sending them over GSM/GPRS networks. Due to limited channel capacity for the codec the DSP with a relatively low computational power is sufficient . Wavelet filters banks are implemented by lifting schemes for a 9-7 and a 5-3 filter. The proposed codec is comparable with a JPEG2000 codec in the rate-distortion sense.

COBISS.SI-ID: 9111318
3.
Transmission of images over general packet radio services using advanced wavelet image codecs and digital signal processor chips.

In this paper we present transmission of static images, acquired by a camera using developed hardware with a DSP TMS320C2407. The GSM/GPRS network was used for transmission of packets using TFTP/UDP protocol. We observed the communications properties and delays caused in the GSM/GPRS network.

COBISS.SI-ID: 11058454
4.
Wavelet-based despeckling of SAR images using Gauss-Markov random fields.

We describe a wavelet-based despeckling algorithm for SAR images. They are first transformed by dyadic wavelet transform (WT). A noise image is modeled as an additive signal-dependent Gaussian noise in WT space and noise-free image with generalized Gauss-Markov random field (GGMRF). For the homogeneous image areas the parameters of the Gaussian distribution and GGMRFs are estimated from incomplete data by mixtures of wavelet coefficients. In heterogeneous areas the unknown parameters are estimated by image and noise models.

COBISS.SI-ID: 11905558
5.
The role of video technology in on-line lectures for the deaf.

Research work represents requirements that disabled people need in frame of interactive video supported user interfaces. Represented system is shown as innovation on this field. This system enables easier integrations in education and qualifying processes of disabled people. Video presentation of information is an important role for information presentation, especially if it’s hard to visually represent. This work was awarded from Award founders board for innovation and innovations training for video online lectures for deaf people.

COBISS.SI-ID: 8933142