Wu Xinhong of Meitu Company: AI technology reshapes the imaging industry

The booming development of smartphones and mobile Internet has promoted the increasing popularity of mobile photography and short video applications, while a new round of AI technology boom has further helped improve the quality and efficiency of the imaging industry. Meitu recently released 7 imaging productivity tools and launched the first large-scale visual model in China that understands aesthetics. Wu Xinhong, founder, chairman and CEO of Meitu, recently said in an exclusive interview with reporters that AI is reshaping the imaging industry, and the value of tools will explode in the AI ​​era. Meitu’s products and services will also be transformed from life scenes. Fully enter the productivity scene.

AI reconstructs the imaging industry and beautiful pictures enter the productivity scene

Over the years, Meitu has always focused on the imaging industry and maintained continued development relying on the extremely popular Meitu Xiuxiu, Beauty Camera and other applications. Recently, Meitu has released 7 imaging productivity tools. These 7 tools are all based on AI and cover visual creation, commercial photography, professional video editing, commercial design and other fields, aiming to comprehensively improve the productivity of the imaging industry.

In the past, Meitu mainly focused on providing services for life scenes, such as shooting, photo editing, video editing and social sharing. In the AI ​​era, in addition to life scenes, Meitu will also provide more productivity scene services. Meitu announced the release of 7 new imaging productivity tools, which will comprehensively expand its products and services from life scenarios to productivity scenarios.

According to Wu Xinhong, Meitu previously established an AI team in 2010 to create the Meitu Imaging Research Institute. The company has a clear development strategy in AI, that is, it will only do business related to vision and will not go beyond this. boundary. In fact, Meitu’s latest release includes modules in four visual application areas: AI pictures, AI videos, AI design, and AI digital people. In the visual field, Meitu is re-developing products and using AI to completely transform core functions. ”

With the launch of 7 imaging productivity tools, Meitu has basically laid out a product matrix in the imaging field, such as Meitu Xiuxiu and Beauty Camera for pictures, and Wink and Start Shooting and Design for videos. Aesthetic design studio and Whee. Next, Meitu’s focus is to continuously verify and iterate on product functions.

For Meitu, Wu Xinhong believes that the current evolving trend of artificial intelligence means that the company needs to reconstruct its products and organization. Meitu's previous products can actually be reconstructed using AI. For example, in computer vision, computer graphics, etc., the original products can no longer meet the current users' needs for better effects, so there is a need to improve existing products. Carry out reconstruction and gradually realize AI-based many core functions.

"In the AI ​​era, the value of tools will explode"

From life scenes to productivity scenes, from apps such as Meitu Xiuxiu and Beauty Camera to imaging productivity tools, this also highlights the current new development strategy of Meitu.

"When everyone mentioned Meitu Xiuxiu before, they thought it was an App that focused on life scenes. But this time we focused on the image productivity tool, which can be said to focus on work scenes, so that everyone has an understanding of Meitu. The impression has changed. Meitu is extending and expanding from life scenes to work scenes. This is actually an increase for us, because the underlying capabilities are similar. The key is to match more needs and use the same capabilities to serve well. More customer groups. Wu Xinhong told reporters that Meitu is undergoing the evolution of tool attributes through the superposition of life scenes and productivity scenes.

According to him, Meitu also wanted to get rid of the tool attribute and explore other popular fields such as social networking in the past few years, but now it no longer wants to blindly diversify because the market space of the imaging industry is large enough. Currently, in the global imaging field, there is a design SaaS company abroad with annual revenue of tens of billions of dollars. However, China's design SaaS industry has not yet produced a leading product, which also indicates that this field will still have broad development in the future. prospect.

"In the past, we were not willing to make tool applications, but now everyone can see that the tool market can also give birth to companies with a market value of more than hundreds of billions of dollars, so now we are also willing to make tool applications." Wu Xinhong said, The value of tools in the AI ​​era will explode. Tool applications can provide deterministic business opportunities. As long as companies have the patience to polish their products, they can make money in work scenarios with a high probability.

He believes that in the future, the market demand for imaging tools will explode, because AI can greatly improve work efficiency, and once users experience the benefits of the tools, they will no longer refuse. At present, the market is in a stage of alternation between old and new ways of working. As more people migrate to AI in the future, productivity is expected to see a significant improvement, and many development opportunities will inevitably arise in this process. Through the close integration of AI and imaging productivity tools, the threshold for creating high-quality content will continue to be lowered, thereby improving the production efficiency of the imaging industry.

The "Battle of Hundreds of Models" must first have a skill, which can be complemented by cooperation with other large models in the future

Since the beginning of this year, generative AI large models have suddenly emerged and become the focus of development in the global technology community. Major domestic technology manufacturers are rushing to launch large model products. Meitu recently released Miracle Vision, the first large visual model in China that understands aesthetics.

Wu Xinhong introduced that Meitu has integrated long-term accumulated aesthetic cognition into large visual models, and built an aesthetic evaluation system based on machine learning to score "aesthetic scores" for model generation results, thereby continuously improving the model's understanding of aesthetics. . In addition, visual models have a unique ecosystem of aesthetic creators, with artists, designers and other professionals with profound aesthetic backgrounds participating in the construction of visual models.

Different from other large models on the market, the large visual model Miracle Vision has strong visual expression and creativity, and reverses technological evolution from visual creation scenarios such as painting, design, film and television, photography, games, 3D, animation, etc. Specializing in Asian portrait photography, national style and fashion, commercial design and other creative directions.

He believes that launching large-scale models is not very difficult. The key is to ensure excellent generation results, which will also test the true technical strength of the company. Meitu will focus on the visual domain and only solve specific problems in specific domains, rather than using general-purpose large models.

Continuous participation is necessary to achieve a good "model ecology", and Wu Xinhong emphasized the importance of creators and developers in this process. In the future, Meitu will establish a model ecosystem to train, distribute and commercialize models, and build them systematically.

In addition, he believes that large models can actually be linked in the future. For example, Meitu is good at visual large models and can cooperate with other large models with stronger language cognitive abilities. "In the current large-scale model boom, companies should have at least one expertise, and Meitu has a clear advantage in the field of visual applications." Therefore, we are also happy to see the "War of 100 Models" and "War of 1,000 Models". In the end, several companies will win, and then we can cooperate with each other to better meet user needs. "Wu Xinhong said.

The above is the detailed content of Wu Xinhong of Meitu Company: AI technology reshapes the imaging industry. For more information, please follow other related articles on the PHP Chinese website!