Video streaming platform iQiyi, also known as China’s Netflix, has been researching the use of facial recognition for animated faces of their shows. In late June, a research team that consists of people from the Chinese streaming site and from Beihang University published a paper introducing an animated character facial recognition dataset called iCartoonFace.
The dataset contains over 400,000 images of more than 5,000 animated characters. It also includes not only human-like figures, but also animals and monsters—which isn’t easier than facial recognition systems from sites like Facebook and Google Photos.
Facial recognition can be a lot harder in animated characters because character can look different depending on the show or universe. For example: Ash in the recent film “Pokémon: Mewtwo Strikes Back—Evolution” uses computer graphics, while the original anime was hand-drawn.
How facial recognition data for animation can help streaming platforms
By having the feature on their streaming platform, iQiyi says that the technology could be used for things like automatic editing, filming, advertisement recommendations and computer-aided modelling. It could also let viewers skip to specific scenes where their favourite characters appear—which is also info that marketers can use as a way to gauge how popular a performer is.
As for why concentrate on animated characters instead of real-life faces on streaming platforms, animated shows are currently thriving. One estimate from Research and Markets estimates the industry in China was worth 200 billion yuan (RM122 billion) last year. The animation industry is also booming due to the COVID-19 pandemic, as it’s a form of entertainment that can be produced remotely.
Although the technology can be used in many different ways, it is still lacking because of uncertain variables that exist like different poses, expressions, illuminations and makeup. Also, with different colours and textures, the process for machines is actually more complex than actual human faces—as cartoon faces they’re often exaggerated.
iQiyi has also been working on other AI projects—including machine learning programs that could help with editing footage or casting actors. The casting could done by matching role descriptors with an actors’ previous performances. However, it’s only used for supporting roles right now.
“Going forward, iQiyi will continue exploring the endless possibilities of entertainment experience by breaking through technological boundaries,” said Liu Junhui, iQiyi’s Senior Director.
[ SOURCE, IMAGE SOURCE ]
