Abstract:

This paper presents several difficulties when teaching signal and image processing courses. Students have to learn broad field of education, from mathematics to engineering. However, it is also important to learn applied mathematical concepts that can be used in real life scenarios. Several concepts that can be learned in one dimensional signal processing can be also applied in two or more dimensions (such as two dimensional images). As a case example, paper describes one laboratory exercise using objective image quality measure application, called "IQmeasure". In application, it is possible to read any image and degrade it using Gaussian noise, degradation using compression, blur degradation and any combination of those degradations, with custom levels of degradation for each one. Afterwards, 18 full-reference image quality measures can be calculated between original and degraded image, for example well-known Peak Signal to Noise Ratio (PSNR) and Structural Similarity (SSIM) index. In this application, it is possible to show that some image quality measures can obtain similar scores for drastically different subjective image quality. IQmeasure can be used in different image, video or signal processing courses. Application itself can be downloaded from http://msl.unin.hr for further educational or research related activities.

Keywords:

Digital image processing, Image quality measures, IQmeasure, Peak Signal to Noise Ratio, Structural Similarity Index.