The development of 5G technology will also bring about a new wave of video technology. 5G technology will provide higher bandwidth and lower latency, which will enable more efficient transmission of video data. This will lead to a new wave of video technology, such as 8K video, virtual reality, and augmented reality.
In conclusion, the law of technology predicts that the technological revolution in the field of video is just around the corner. With the development of AI and 5G technology, the video industry will experience a new wave of innovation and development.
As the popularity of ChatGPT grows, Google’s new AI model Phenaki has been generating buzz for its ability to produce up to two minutes of story-telling video in just minutes with a single cue word. The video generated by Phenaki is longer and more sophisticated than its predecessors, with a high resolution and an understanding of different art styles and 3D structures. Additionally, Google has launched Imagen Video based on the diffusion model, which also boasts high resolution and the ability to understand different styles. As big model technology continues to evolve, the next breakthrough point in AI is eagerly anticipated.
Meta has also joined the race of video generation with its Wensheng video model, Make-A-Video, an upgraded version of the text image generation model. The AI model has three parts: Text Image Generation Model P, Spatio-Temporal Convolutional Layers and Attention Layers, and Frame Interpolation Network to increase frame rate and two super-resolution models to improve image quality. Make-A-Video can convert still images into videos, create a video based on two pictures before and after, and generate a new video based on an original video. As the competition in big model technology heats up, the potential for new breakthroughs in video generation is highly anticipated.
VidPress is an intelligent video synthesis platform developed by Baidu in China. It uses Wenxin large model to automate the five steps of text script, video content search, material processing, audio and video alignment, and editing. AI-generated video content is becoming increasingly popular in the fields of cultural entertainment, education, media, and more. In 2022, high-quality Vincent graph models such as DALL E, Imagen, and Stable Diffusion will help to break through the commercialization bottlenecks of AI-generated content, such as difficulties in monetization. VidPress is a great tool to help create professional-level AI-generated video content.
As companies compete in the rapidly evolving field of big model technology, there are two types of companies worth paying attention to: those with data foundations and application scenarios, and those with revolutionary technology reserves. Large industry players like Netflix and Disney have a distinct advantage, with access to billions of member evaluations and data on viewer habits and needs. Netflix has even used AI algorithms to generate movie thumbnails based on user preferences, and recently released an AIGC-animated short film “Dog and Boy,” with animation scenes created by AI. Additionally, startups founded by core teams of established tech giants, like Image Video, have great potential for growth, with a first-round valuation of $100 million. As the industry continues to develop, opportunities abound for those who are prepared.
Douyin and Kuaishou, two giants in China’s short video industry, are expanding their capabilities with AI technology. Bytedance’s Jianying video editing tool and Kuaishou’s Yunjian both allow users to convert pictures and text into videos by entering keywords or paragraphs of text. Meanwhile, companies with technical reserves on vertical subdivision tracks are also worth watching, such as Moviebook Technology, which specializes in intelligent visual technology. Moviebook Technology released the AGC intelligent image production engine in 2018, based on generative AI technology. These companies have great potential for growth as the industry continues to evolve and expand, with AI technology becoming increasingly important for video production and editing.
Moviebook Technology’s AGC solution uses MCVS technology to automatically understand and label various content in videos, and then uses computer image technology to generate brand new videos through the video content automatic production engine MAPE. This technology incorporates artificial intelligence multi-modal semantic understanding and uses deep learning to achieve video automation processing and synchronous generation of video, innovating the way of video image production. AGC can generate personalized video content in a short time and reconstruct shot videos, including automatically anchoring key frames and generating content that was not in the original image. According to public data, the total cost of a 60-second video generated by Moviebook Technology’s AI generation engine is reduced by over 79.8% compared to traditional methods, while productivity can be increased by over 100 times. AGC is mainly used in government services, enterprises, science and education, pan-entertainment, media, cultural tourism, and other fields.
The growth of the artificial intelligence generative content (AIGC) industry is accelerating due to the popularity of large models such as ChatGPT, but Gartner’s 2022 Emerging Technology Maturity Curve shows that the industry is still in its infancy. However, as more value is placed on technical bases such as large models, the industry is showing an acceleration in growth. There is a lot of expandable space for AIGC in other vertical fields, and it is predicted that the AIGC market size will exceed one trillion yuan by 2030. To seize this opportunity, the Chinese industry needs to pay attention urgently. Seed players on vertical tracks, such as Movie Book, have already received investment from State-owned Assets Funds, and set a new financing record of CNY 1.36 billion in the field of AI visual production.