Implementation and evaluation of discrete wavelet transform based image coder

U okviru ovog završnog rada ostvaren je transformacijski koder primjenom diskretne wavelet transformacije (DWT) kojom se provodi višerezolucijska analiza slike. Omogućena je promjena vrste waveleta, stupnja kompresije i broja dekompozicija slike. Opisana su svojstva DWT-a, proces kvantizacije te statističko kodiranje. Objektivnim mjerama za određivanje kvalitete slike – PSNR, VSNR, WSNR, SSIM i UQI, programsko je rješenje evaluirano te je prikazana statistička obrada rezultata. Dobiveni rezultati uspoređeni su sa slikama oštećenim JPEG2000 koderom iz baze slika TID2008. Također, napravljena je i usporedba s koderom koji rabi diskretnu kosinusnu transformaciju (DCT).Within the scope of this final paper, a transformation encoder was implemented using discrete wavelet transformation (DWT), which enables multi-resolution analysis of images. The ability to change the type of wavelet, compression ratio, and number of image decompositions was provided. The properties of DWT, quantization process, and statistical coding were described. Objective image quality assessment methods, such as PSNR, VSNR, WSNR, SSIM, and UQI, were used to evaluate the software solution, and statistical analysis of the results was presented. The obtained results were compared with images distorted by the JPEG2000 encoder from the TID2008 image database. Additionally, a comparison was made with an encoder that uses discrete cosine transformation (DCT)