With the advent of 5G wireless communication systems, multimedia data is predicted to grow rapidly in the near future. As a result, the large amount of multimedia data puts huge pressure on wireless communication, which poses a huge challenge to network capacity. Multimedia data (image/ video) intelligent compression is an effective way to increase the network capacity and improve the user's QoE. The success of image/video compression generally depends on data representations. Although specific domain knowledge can be used to help design representations, learning with generic priors can also be used, and the quest for AI is motivating the design of more powerful representation learning algorithms implementing such priors. This article proposes an intelligent computing communication framework to reduce the amount of transferred data. Simultaneously, it reviews the area of representation learning, covering advances in dictionary learning, RoI, and deep learning image/video compression methods. Furthermore, we compare various compression methods as well as point out promising future works.

中文翻译：

---

无线多媒体通信中的表示学习

随着5G无线通信系统的出现，预计多媒体数据在不久的将来会迅速增长。结果，大量的多媒体数据给无线通信带来了巨大的压力，这对网络容量提出了巨大的挑战。多媒体数据（图像/视频）智能压缩是增加网络容量和改善用户QoE的有效方法。图像/视频压缩的成功通常取决于数据表示。尽管可以使用特定领域的知识来帮助设计表示形式，但是也可以使用具有通用先验的学习，并且对AI的追求正在激发设计实现此类先验的更强大的表示学习算法。本文提出了一种智能计算通信框架，以减少传输的数据量。同时，它回顾了表示学习的领域，涵盖了字典学习，ROI和深度学习图像/视频压缩方法方面的进展。此外，我们比较了各种压缩方法，并指出了有希望的未来工作。