Founded in 2014, Meishe Technology is a national high-tech enterprise focusing on the overall solution of intelligent audio and video. In the field of video and audio for many years, Meishe Technology, with its strong technical strength, continues to expand application scenarios under the current wave of video, and extends services to many fields such as smart phones, smart hardware, smart cars, smart radio and television, helping the platform to serve users. Bring a richer and more convenient video creation experience. This time, LiveVideoStack had the honor to interview Hou Kang, an algorithm engineer from Meishe Technology. He introduced to us the focus of Meishe Technology in the audio and video field in the past year, and what it has done in the development process of Meishe's car intelligent editing solution. The "pits" encountered in the process of work and research and development, and the recently launched desktop-side in-depth video editing solution "Meiying PC Edition", etc.
Hou Kang: Beijing Meishe Network Technology Co., Ltd., Algorithm Engineer
LiveVideoStack: Hello, Mr. Hou. Please introduce your work in Meishe Technology and your recent technical direction.
Hou Kang: I am currently in charge of the management of the Meishe AI algorithm team. The technical aspects are mainly for the development of intelligent editing and image processing algorithms. At present, I mainly focus on some new AI research directions, such as the application points that transformers may involve in the intelligent editing system, and at the same time continue to expand and optimize the intelligent editing capabilities of Meishe.
LiveVideoStack: In the past year, what is the focus of Meishe Technology in the field of audio and video? What results have been achieved?
Hou Kang: In the audio and video field, Meishe mainly focuses on the expansion and application of intelligent editing in various fields. Meishe developed an intelligent editing system for life and travel videos in 2019, and received a warm response from the market. Therefore, we have carried out a lot of optimization work for various scenarios in the future. For example, in the field of smart cars, we have done a lot of adaptation and optimization work for the characteristics of car video shooting scenes and the hardware system of the car, and added characteristic algorithms for cars, which have received unanimous praise from customers.
Last year, in addition to continuous efforts in the fields of smartphones, smart hardware, social platforms, government media, and overseas, Meishe also expanded its service areas to new industries such as smart wear and smart cars. In terms of technology, we have also made new breakthroughs. In the first "Radio, TV and Network Audio-visual Artificial Intelligence Application Innovation Competition" held by the State Administration of Radio and Television, Meishe's intelligent video content production platform based on the deep learning framework has been selected and selected. Won the first prize.
LiveVideoStack: Congrats here! We are also delighted to find that Meishe has begun to develop in the field of smart cars. What is the opportunity to develop this scene?
Hou Kang: With the popularization of vehicles, hardware upgrades and the improvement of users' cognition, car users' demands for vehicles have also changed from basic driving to a comprehensive driving experience. Simple in-car entertainment can no longer meet user needs. The sharing of life content in the car is becoming more and more important. Most of the current shooting hardware still uses external devices such as mobile phones and cameras, which is not convenient for users to use during driving, and the shooting angle is also extremely limited. In addition, there are very few production tools suitable for vehicle scenes at present. When users use mobile phones and PC software for post-processing, they need to import long videos. Post-production is difficult and cannot be shared in real time. In this context, Meishe Automotive Intelligent Video Processing Solution came into being.
LiveVideoStack: As a new video production platform, what is the difference between video processing in automobiles and video creation through mobile terminals, PC terminals, etc. in Internet scenarios?
Hou Kang: The car itself does not have a convenient user interface like mobile phones and PCs, so the need for automated editing solutions is more urgent. The processing platforms and systems of automobiles are still in full bloom, so a lot of adaptation work needs to be done to meet the needs of different manufacturers. Meishe currently has the ability to adapt to all platforms. In addition, cars cannot update software at any time like mobile phones and PCs, and have higher requirements for safety and stability. Meishe has done a lot of work to meet the requirements of car manufacturers for stability and safety. At present, our SDK and the whole set of intelligent editing system can stably run on the car operating systems of mainstream manufacturers.
LiveVideoStack: What are the "pits" encountered during the development of the intelligent video production platform in the smart car scenario? How to solve?
Hou Kang: In the development process of the intelligent editing system, the main difficulty lies in how to optimize for different scenarios. At present, AI technology cannot complete the editing of different scenes and different types of films like a real director. Therefore, different frameworks and algorithms need to be designed for different scenes to achieve the best effect. In addition, when developing a car intelligent editing system, the biggest problem is the adaptation process of different systems and hardware architectures. Since automotive hardware, processing platforms, and systems are not as unified as mobile phones and PCs, a lot of work needs to be done for adaptation and tuning in order to achieve smooth and stable operation.
LiveVideoStack: I saw that Meishe recently launched the desktop-side in-depth video editing solution "Meiying PC Edition". In fact, there are many video editing tools for content creation, whether it is PC or mobile, so how did Meishe consider launching such a product?
Hou Kang: Meiying PC version is an in-depth video editing tool for enterprise users. It has comprehensive and efficient professional editing functions, and the access method is simple. It supports enterprises to add self-developed or other third-party technologies. Based on the cross-platform Meishe SDK, enterprises can build interoperable and collaborative platforms and solution products, and output multi-scene professional editing application products.
The launch of this product by Meishe is inseparable from the current development trend in the video field. Nowadays, with the further intensification of competition between long and short video platforms, medium video content has become the focus of the platform. Medium and long videos have stricter requirements for creation tools and scenes, and the style and tone of different platforms also determine that it is difficult for creators to directly produce the required content through traditional video editing tools such as Pr and Final Cut Pro, and create their own Video editing tools are the choice of many platforms. However, the development of video and audio editing on the PC side is difficult, and the required investment and post-maintenance costs are too high. Based on this background, we launched the PC version of Meiying to help the platform quickly build its own video creation platform and build a systematic product ecosystem.
LiveVideoStack: What is the next step of Meishe Technology? Will the focus be on technology research and development, or on scene expansion?
Hou Kang: As a national high-tech enterprise, Meishe is committed to becoming the most influential intelligent video and audio technology enterprise in the world. Meishe's core R&D team has rich R&D experience in the field of video and audio, focusing on the development of video and audio engines, graphics and image processing and computer vision for more than 20 years. At the same time, we are constantly exploring new application scenarios. Last year, we successfully expanded our business to smart cars, smart wear and other industries on the basis of smart phones, smart hardware, radio and television and communication operators.
In our opinion, technology development and scenario expansion are mutually reinforcing relationships. The leading technical level can support us to develop a wider range of application scenarios. At the same time, new scenario requirements, new situations and new problems encountered in the application will also prompt us to continuously improve our technical capabilities. This year, we have launched a variety of new solutions such as the PC version of Meiying and one-click filming, which can adapt to more application scenarios, so stay tuned.
LiveVideoStack: What exciting content will you share at this conference? Can you give me some spoilers?
Hou Kang: At this conference, I will share the solutions related to Meishe's automotive intelligent editing system, including image quality enhancement, intelligent video editing and virtual scene entertainment.
LiveVideoStack: Okay, thank you for the interview, looking forward to your wonderful sharing at the conference.
Interview, Editor: Teresa
▼Scan the QR code below to learn more about the Audio and Video Technology Conference▼