DESIGN NAME: Spectator Assistant
PRIMARY FUNCTION: Video Application
INSPIRATION: As people pay more and more attention to sports events, audiences want to enhance their viewing experience and get more professional information. The audiences lack knowledge of the events, so we hope to provide the popularization of relevant knowledge. And they can't see their favorite players clearly during the game, so we hope to provide path tracking and point-by-point reviews. Color blind and hearing impaired users cannot effectively access information, so we hope to provide color vision correction and subtitles.
UNIQUE PROPERTIES / PROJECT DESCRIPTION: Spectator Assistant is a product used for watching sporting events. It can identify and analyze information related to the games, such as real-time positions, personal information, highlight moments, technical terms, then visually present them to reduce users' confusion. The lively AI digital human can improve users' emotional experience. Spectator Assistant can improve color recognition for color blind people. And through Intelligent Subtitles, the hearing-impaired audience can enjoy the game equally. It can also help users quickly generate video clips.
OPERATION / FLOW / INTERACTION: When identifiable information appears, the AI Automatic Recognition function guides users to identify players or objects through dynamic effects and vibrations, then the interface displays relevant results. The process will not interfere with the user to watch the video. And set the Track Tracking function to select the players to follow, and you can see the position of the players in real time. In addition, Color Vision Assistance and Intelligent Subtitles can help color impaired and hearing impaired users get clearer watching effects. And you can select the designated player or action tags, quickly edit the video material by the Smart Clip.
PROJECT DURATION AND LOCATION: The project started in April 2022 in China and and is continuing.
|
PRODUCTION / REALIZATION TECHNOLOGY: The product uses Knowledge-based Video Semantic Understanding technologies, such as Voice Recognition, Face Recognition, and Body Recognition, to parse and merge multi-modal content such as voice, vision, or text, so as to realize intelligent retrieval and analysis of event information. The use of Video Color Vision Assistance technology can ensure the contrast of video screen colors for color impaired users. Intelligent Video Content Production technologies, such as Video Stripping and Intelligent Editing are used to aggregate video scenes and present wonderful event moments.
SPECIFICATIONS / TECHNICAL PROPERTIES: Mobile phones, computers, TVs and other multi-terminal devices are supported at the same time.
TAGS: sports, events, information visualization, barrier-free, video
RESEARCH ABSTRACT: Through questionnaire survey and user interview, we have had an in-depth understanding of the needs and pain points of the spectators. The survey have found that the general audiences lack of understanding of the events, professional audiences can't clearly find or see their favorite players in the competition, and they are eager to share after the competition. At the same time, we have found color blind and hearing impaired users cannot effectively access information while watching the events. In addition, we have also studied the problems that can be solved by the existing Video Technology and AI ability.
CHALLENGE: The challenge is how to present events to users in a visual way. And for that we have explored different ways of perception and triggering of information recognition results, and have used emotional expression such as the digital human and dynamic effect in different scenarios. In addition, we have combined the AI technology of Video Content Analysis, Color Vision Correction and Intelligent Content Production to make the game process more intuitive.
ADDED DATE: 2022-09-30 03:03:07
TEAM MEMBERS (6) : Jia Li, Yuqian Zhang, Juan Xiong, Qiianle Wu, Tingting Wang and Yu Youping
IMAGE CREDITS: Copyrights belong to Baidu Online Network Technology (Beijing) Co., Ltd, 2022
PATENTS/COPYRIGHTS: patented.CN202230477563.5. Baidu Online Network Technology (Beijing) Co., Ltd, 2022
|