Analog Joint Source Channel Coding for Image and Video Encoding and Transmission

Resumen

The use of Internet-connected devices, particularly smartphones and Internet of things (IoT) devices, has significantly increased in recent years. The latter are experiencing a tremendous growth thanks to lower costs, improved wireless communications and new applications. Although there are countless applications, it is the multimedia applications that make use of images and video that require the most computing power and bandwidth resources. Today, the most advanced image and video transmission systems focus on digital solutions that require complex encoding systems with many processes to achieve high compression rates without losing visual quality. In addition, digital transmission systems need to introduce redundancy to avoid errors caused by the channel, require a feedback channel between the receiver and the transmitter to adapt the transmission rate to the channel conditions and, in case these mechanisms fail, the information must be retransmitted to allow its correct decoding. In this work, analog image and video encoding and transmission systems are proposed focusing on IoT devices with limited data processing resources and especially considering broadcast wireless communications. The proposed systems adapt the received visual quality to the channel conditions without channel knowledge. The most relevant information for visual quality is protected against channel distortions by means of redundancy, while the total amount of transmitted information can be adjusted to the available bandwidth and the desired visual quality by discarding certain information to be compressed. This work contains a detailed review of the state of the art, including an analysis of the main signal processing tasks and mathematical transformations used. Considering this analysis and the scenario requirements, several encoding and transmission schemes have been proposed for multimedia content which are based on the use of analog joint source-channel coding. First, all design decisions with respect to the proposed image and video systems are analyzed, and then their performance is experimentally evaluated according to image quality metrics such as the structural similarity index measure (SSIM) or the peak signal-to-noise ratio (PSNR). Subjective analyzes of perceived image quality are also performed. In all cases, the results obtained are contrasted with those produced by reference systems available in the literature, including analog, hybrid, and fully digital approaches. In the case of the proposed system for analog image coding and transmission, the evaluation is carried out not only considering simulated channels, but also over-the-air transmissions performed with a testbed.