A Survey on Generative AI and LLM for Video Generation, Understanding, and Streaming

Pengyuan Zhou,Lin Wang,Zhi Liu,Yanbin Hao,Pan Hui,Sasu Tarkoma,Jussi Kangasharju
2024-01-30
Abstract:This paper offers an insightful examination of how currently top-trending AI technologies, i.e., generative artificial intelligence (Generative AI) and large language models (LLMs), are reshaping the field of video technology, including video generation, understanding, and streaming. It highlights the innovative use of these technologies in producing highly realistic videos, a significant leap in bridging the gap between real-world dynamics and digital creation. The study also delves into the advanced capabilities of LLMs in video understanding, demonstrating their effectiveness in extracting meaningful information from visual content, thereby enhancing our interaction with videos. In the realm of video streaming, the paper discusses how LLMs contribute to more efficient and user-centric streaming experiences, adapting content delivery to individual viewer preferences. This comprehensive review navigates through the current achievements, ongoing challenges, and future possibilities of applying Generative AI and LLMs to video-related tasks, underscoring the immense potential these technologies hold for advancing the field of video technology related to multimedia, networking, and AI communities.
Computer Vision and Pattern Recognition,Artificial Intelligence,Multimedia
What problem does this paper attempt to address?
The paper aims to explore how Generative AI and Large Language Models (LLMs) are reshaping the field of video technology, including video generation, understanding, and streaming. Specifically, the paper focuses on the following aspects: 1. **Video Generation**: The paper examines innovative applications of Generative AI in creating highly realistic videos, which can significantly narrow the gap between real-world dynamics and digital creations. 2. **Video Understanding**: It investigates the advanced capabilities of LLMs in video understanding, such as extracting meaningful information from visual content, thereby enhancing our interaction with videos. 3. **Video Streaming**: It discusses how LLMs contribute to a more efficient and user-centric streaming experience by adapting to individual viewing preferences and optimizing content delivery. The paper not only reviews current technological achievements but also highlights the challenges and future research directions, emphasizing the immense potential of these technologies in advancing the multimedia, networking, and AI communities within the video technology field. Additionally, the paper underscores the importance of responsibly developing and using AI in video technology, balancing technological innovation with ethical considerations.