Performance comparisons of the image quality evaluation techniques in Wireless Multimedia Sensor Networks

Recent developments with the integration of new vision capabilities into the wireless sensor devices have resulted in the necessity of investigating the behavior of well known image quality evaluation techniques in Wireless Multimedia Sensor Networks (WMSN). Wireless sensor devices can collect process and evaluate image data with a camera module. These devices can also monitor the image quality in order to take some adaptive precautions at the communications layer protocols and thus guarantee application layer quality requirements. In order to evaluate the transmitted image quality results more accurately, it's imperative to analyze the factors which lead to change in behaviors of existing image quality evaluation techniques. In this work, the effect of the transmission distortions, packet size, link and node correlation on the image quality evaluation are highlighted. The success of prevalent image quality metrics is also examined with emphasis on whether these metrics can reflect implicitly the region of interest (ROI) information in an image. Furthermore, a simple metric is proposed to estimate an image quality on the fly by considering the ROI parts of the image and using only the average of the packets and their priority level. This study leads to explore smart ways for the adaptation of the communication protocols to the wireless environment.