People's Daily: Video Cloud AI Editorial Department empowers People's Daily to report in all media.

Alibaba Cloud AI Editorial Department helps People's Daily to improve the quality and efficiency of media content production based on big data and artificial intelligence technology.

People's Daily is the newspaper with the largest circulation in China, and 1992 has been rated as one of the top ten newspapers in the world by UNESCO. Over the past 70 years, People's Daily has actively publicized the Party's theory, line, principles and policies, actively publicized the major decisions and arrangements of the central authorities, and timely disseminated information in various fields at home and abroad, making important contributions to the unity of China and leading the people of the whole country to win great victories in revolution, construction and reform.

During the two sessions, People's Daily actively promoted the implementation of AI technology and empowered all-media production from multiple directions and angles, hoping to improve the production efficiency and communication effect of news content through intelligent production methods. The main requirements are:

In the layout process of all-media communication matrix such as newspapers, periodicals, websites, terminals and micro-screens, various media carriers and channels undertake different media content production. How to make reporters quickly search for news materials from all media channels in a short time, and how to make editors quickly edit and release a wonderful golden speech video are the keys to achieve a creative breakthrough in the reporting experience of the two sessions under the background of high business volume during the two sessions.

In order to achieve a wider and deeper news communication effect, it is an important means to build an all-media communication matrix to enhance audience coverage. Generally speaking, products with strong technological innovation have better expressive force and are more easily accepted by the audience. Then, according to the audience of different channels, CCTV needs to create suitable, rich and wonderful content to enhance its communication power.

The National People's Congress and the National People's Congress are important and special national events, during which the censorship of media content needs to be treated very seriously and strictly. For example, sensitive and illegal contents such as politics, pornography, violence, terror and pornography need to be reviewed to avoid violations or political mistakes. However, review is a rigorous and time-consuming task. How to identify illegal content accurately and efficiently, so as to improve the quality of review and reduce labor costs?

The AI editorial department provided by Alibaba Cloud Video Cloud's intelligent media production products is an intelligent content production line for new media. With the help of big data technology and artificial intelligence, we can efficiently handle massive videos, pictures and texts that meet the news business, realize the automatic, batch and intelligent production of video manuscripts and graphic manuscripts according to the set themes, enrich the manuscripts and resource applications of media users, and support the core business of media users. AI editorial department empowers CCTV to be faster, better and wider. Core competencies include:

Through the deployment of Alibaba Cloud private network VPC, the system can flexibly expand capacity according to business requirements, and can seamlessly expand cluster resources when dealing with high concurrent requirements. At the same time, it can be combined with GPU server to provide more powerful video processing and computing support to meet the higher stability requirements during the two sessions.

1. The smart tag function can intelligently identify various news scenes such as speeches, press conferences, group photos and meetings. By understanding the dimensions of video such as labels, subtitles and voice, unstructured video can be transformed into structured information such as labels, which is convenient for editors to find all-media materials more accurately and quickly, and comprehensively improve the efficiency of news gathering.

2. The cloud desktop fine editing function integrates mainstream non-editing tools such as Premiere, ocean non-editing and new automatic non-editing, and realizes faster calculation, larger storage and more flexible scheduling than local ones. In the news scene, reporters can enjoy the smooth editing experience supported by remote high computing power by simply linking to the non-editing high-performance server on the cloud with one click, and even the 4K scene can be edited in a short time, which greatly improves the editing efficiency.

3. The simultaneous subtitle function supports the rapid generation of subtitles matching the video according to speech recognition, and supports the replacement of subtitles in batches, which can improve the efficiency of video production for new media interviews and conference scenes.

4. The intelligent watermark and subtitle function supports clearing all kinds of watermarks and subtitles preset in the video, restoring the pictures blocked by watermarks and subtitles, realizing intelligent image filling and improving video processing efficiency.

1. The function of changing horizontal screen to vertical screen supports automatic and comprehensive analysis of characters and scenes in horizontal video, selection of appropriate pictures, and production of vertical video. At the same time, the integrity of key information is maintained, and the focus on the core content of the video is realized, which is convenient for newspaper editors to publish works on short video platforms, and can also meet the audience's viewing needs for a single character and the viewing habits of mobile terminals.

2. Based on the rich video AI capabilities, the intelligent striping function is combined with the editing function, and the video is intelligently striped in different dimensions such as the character dimension and the scene dimension, so that the intelligent striping of news videos can be realized, and more independent materials can be efficiently produced.

3. With the help of AI capabilities, the intelligent poster function is based on the poster template library in Alibaba Cloud, covering a large number of topics such as current political reports, party building propaganda, poverty alleviation, and national anti-epidemic. According to the editing requirements, you can quickly search for templates, replace pictures, generate posters with various themes, and help the People's Daily digital screen to promote new media, popularize policies, and broadcast current political hotspots.

4. The data news function makes editing video as convenient as editing tables through Excel by converting a large amount of data into visual information, and enhances the expressive force of news content through this creative and intuitive information way, thus improving the communication effect.

The intelligent audit function relies on the official massive public opinion early warning big data platform, and with the help of mature artificial intelligence algorithms, it can accurately identify political figures appearing in videos and pictures; Second, it can automatically detect fallen officials and inferior artists; Third, it can intelligently judge pornography, violence, weapons, horror, picture-in-picture and four illegal scenes, meet the standards and requirements of newspapers for content review, provide reference opinions for subsequent manual review, greatly reduce the cost and time of manual review, and effectively avoid the risk of adverse political impact caused by manual review omissions.

Through the deployment and application of Alibaba Cloud video cloud software and hardware cluster resources and a number of AI functions, the coverage of People's Daily, Internet, terminals, micro and screen media has been greatly improved in terms of quality, efficiency and communication effect.

Multimedia AI(MultiMedia AI) is a general technology platform product of video AI jointly developed by Video Cloud and Dharma Institute. Combining vision and multimodal technology, we can realize the basic functions of core video application fields such as video structuring, video fingerprint search and video production, and provide customers with the core capabilities of video content understanding, copyright protection, video cataloging, video search recommendation, video interaction and video-assisted production, so as to improve personnel efficiency and traffic efficiency.

For more information about multimedia AI, please refer to the product details page of multimedia AI.

Original link:/document _ detail/310955.htm? UTM _ content = g _ 100032958 1

This article is the original content of Alibaba Cloud, and may not be reproduced without permission.