Smartphones have developed to this day, and AI technology has penetrated into many fields such as display, imaging, and gaming, becoming an important part of the competition for flagship chip product strength. Recently, at the Dimensity flagship technology communication meeting held by MediaTek , AI Image Semantic Segmentation has attracted a lot of media attention. It can process images accurately and efficiently, which can not only greatly reduce computing power requirements, but also take the results into account, which will inevitably bring the "volume" of mobile phone images and display applications to a new level.
As we all know, MediaTek ranks first in the global smart TV chip market for many years. Its smart TV chips are highly recognized by TV manufacturers and consumers due to its strong performance and technical advantages such as AI picture quality enhancement technology (AI-PQ).
is the deep accumulation of AI technology in the field of smart TVs, which gives MediaTek a competitive advantage in AI image processing capabilities. MediaTek is the first to bring the latest flagship TV technology to the mobile platform, helping smartphones to "not only be used for a long time" and "run fast", but also be able to "shoot beautifully" and "see beautifully".
In terms of smartphone display, users have a strong demand for the image quality of smartphones. The traditional image processing method is time-consuming and labor-intensive and cannot effectively handle complex scenes and characters. How to present images that users are satisfied with is a problem that mobile phone chip and terminal manufacturers are currently to be solved.
MediaTek proposes to use the advanced image enhancement technology in flagship TV chips to use the camera and display functions of flagship mobile phone chips to comprehensively improve the visual and capture experience. At the same time, with the help of AI image semantic segmentation technology, powerful object recognition and precise area segmentation capabilities, it can not only reduce mobile phone power consumption but also ensure visual effects.
AI image semantic segmentation technology can perform differentiated image processing for different object characteristics of complex scenes to create a more three-dimensional depth of field effect. The picture scene can also be judged by the combination and proportion between image objects, and targeted processing around the characteristics of the scene to achieve image optimization.
For example, the blue sky, green space, straw mat, portrait and other elements appearing in the picture are quickly classified and cut through AI image semantic segmentation, and targeted optimization is carried out, thereby presenting a picture effect closer to the reality, allowing users to gain a more delicate and realistic image experience.
has entered the 5G era, and AI multimedia has become the mainstream application, bringing users more abundant multimedia creation methods such as photography, video recording, and live broadcast. The improvement of video quality and effect has become a necessity. At the same time, the increase in video specifications also means that the system burden will be raised, and the mobile phone will be hot. As the temperature rises, the superimposed algorithm is limited, resulting in a certain frame in the entire video being less effective than taking a direct photo.
In the future, through AI image semantic segmentation technology, only the image quality enhancement or special effects processing of this part can be done for different objects, reducing the required computing load, thereby superimposing more optimization algorithms to quickly improve the video recording effect.
AI image semantic segmentation technology can perform differentiated image processing for different object characteristics in complex scenes, and can also generate special effects for subject background objects. In addition, this technology has significant advantages in video beautification processing. Through AI scene quality enhancement technology, the scene content in the image can be recognized in real time to optimize the contrast, sharpness and color of each frame of the picture in scenes such as architecture, landscape, and portraits. On the other hand, the AI area image quality enhancement technology can also perform object recognition and cutting in real time, adjusting the image parameters of different objects in each frame of the picture.
In addition to display, AI image semantic segmentation technology can also be used in the image application of smartphones, making focus catching more accurate. Compared with the traditional focus-chasing method, the subject and background can be automatically distinguished through AI image semantic segmentation technology, and the shutter parameters can be accurately set for the subject to ensure accurate focus and make the capture clearer.
AI is an indispensable core technology for the development of smartphones, and AI multimedia is the mainstream development trend.If high-energy-efficient AI (APU) is MediaTek's first step in laying out mobile phone AI, introducing AI image semantic segmentation technology into smartphones is the second starting point of its innovation.
At this year's MediaTek Dimensity Flagship Technology Communication Conference, we can see that MediaTek once again brought surprise innovations in AI technology, understands users' pain points and needs, explores the application and exploration of AI image semantic segmentation technology, and demonstrates MediaTek's accurate judgment and forward-looking layout of the industry's technological evolution trends. We look forward to the display of this technology's strength on the Dimensity Flagship.
site, MediaTek also shared topics including mobile ray tracing, mobile GPU efficiency enhancement solutions, 5G new dual-pass, Wi-Fi 7, high-fidelity Bluetooth audio, high-precision navigation, etc., leading the innovation trend of mobile platforms with advanced technology and promoting the continuous upgrading of mobile terminal user experience.
