NEW YORK, NY / ACCESSWIRE / January 12, 2020 / Significantly, by 2021, the global media entertainment industry is estimated to exceed a worth of USD 2.6 trillion, to which China will contribute over USD 1.5 trillion. The penetration rate of smart image technology will rise from less than 1/10,000 to 6/1,000. In particular, the aforementioned industry refers to an omni-media entertainment matrix which includes news coverage, sports, dramas and comedy shows, games and ads, radios and podcasts, TV and films, and publishing.
Over the past decade, computer vision technology has been widely applied in Hollywood, partially thanks to the rapidly advanced standardization of technologies and applications. Although the media entertainment industry was under the pressure of transformation in the past year, this trend shows no signs of abating. TV, DVD and CD-ROM are substituted by distribution channels such as streaming media, podcasts, games and online services, resulting in fragmentation of target audience and thereby the need to meet small target groups’ demands in a cost-effective manner. It is expected that this trend will be reinforced by technology, development scale, investment, and market competition in the next few years.
The transformation of technologies and customer behaviors gradually enables media interaction to become more intelligent. Computer vision technology has effectively penetrated all links and stages, including intelligent production of entertainment products (contents) and personalized delivery (distribution and transmission) of finished products.
Content production: Computer vision has become a general technology in the industry, and will shake off the shackles of scale in 2020
High-quality hardware and edge computing vision algorithms have further improved the integration of image capture, which satisfies varying preprocessing requirements (gamma and color matrix) by enhancing stability, reducing noise and eliminating defects. The main task of computer vision is to process captured images or videos to obtain 3D information, keyframe extraction and associated frame association of corresponding scenarios. These applications have been widely applied to variety shows, sports events, and streaming media content production. Current leading computer vision technology appliers are Arcsoft and Moviebook (China), Vinten, A&C, Radamec and MRMC (UK), Panther (Germany), etc.
In addition to image data extraction, concept information (such as speed, distance and 3D shape) can be extracted from any image sequence. Statistics from most of the enterprises involved in the survey show that 3D data acquisition systems are employed by more and more companies to create models for animation in programming and video games. For example, Germany’s X-IST makes it possible to process animation based on facial movements; Moviebook’s MCVS (Motion Capture from Video System) realizes automatic tracking, distance and speed estimation, panoramic reconstruction, etc.
In terms of data synthesis, it has been technically feasible to synthesize background images in real time and to match them to camera viewing parameters (such as position, zoom, and depth of field). This industrial demand allows much more complex visual sets, including ones where there is interaction between roles and synthetic objects. This information is currently obtained through one of two ways: using optical or mechanical sensors fitted to a remotely driven robotic camera; using pattern recognition in the video and then to determine the position.
With regard to content cataloging, post-production workstations will benefit from more automated and adaptive approaches, such as systems from Quantel (UK), Discreet Logic Alias and Wavefront (Canada), and Moviebook (China). These systems have introduced 3D reconstruction, keyframe extracting, filtering, natural semantic understanding, intelligent segmentation, interpolation, etc. Operators supervise the process running over a sequence nearly in real-time, and tunes parameters to secure a satisfactory result.
The effective management of film and video footage is also an important issue for a number of broadcasters and producers. Through field survey, it is found that computer vision technology enjoys unparalleled advantages and has been capitalized on by many companies. For instance, IBM media system indexes content based on color, texture or shape queries; Kwai and Tik Tok realize content-based retrieval based on video frame annotations; Moviebook completes automated annotation of video archives. Suppliers such as Digital Vision (Sweden), Aurora and iQIYI are working on interactive processing of low-resolution content to achieve motion compensation, color correction, scratch, dirt, and noise filtering, etc.
Content delivery and operation: Computer vision technology is playing an irreplaceable role
In respect of commercialization, computer vision has become a mature technology for ad spot scanning and ad insertion in continuous frames. For example, when broadcasting sponsored events such as football matches, motor racing and variety shows, an appropriate video scene can be found for seamless ad placement. Systems able to replace or insert such panels in real time using image processing are on offer from technology leaders including Matra Datavision (France), Orad (Israel), VDI (USA), and Moviebook (China). Resulting additional advertising revenues have become an important business portfolio of TV and streaming media platforms.
The popularization of interactive and personalized technologies facilitated the boom of customized content in 2019. Technology leaders are exploring solutions based on AI and machine learning – surpassing the current recommendation engine (content distribution engine) to provide highly personalized products with content experience at its core, such as interactive dramas, content scenarios produced based on AI, and even customized roles, virtual hosts and roles. One case in point is the virtual host technical solution launched by Moviebook, which enables images to interact with synthetic characters, aligns the two in space and synchronizes them in time. Such interactive applications are applied not only to the media entertainment industry, but more to the education and VR industries.
View source version on accesswire.com: