untitled design

Megogo will launch a neural network to search for mat and smoking scenes in its content

Viktor Chekanov, CEO of Megogo, told Vedomosti that the online cinema had come up with a way to quickly recognize prohibited content. For this purpose, the video service within the next month will create a special program based on a multitasking neural network that will automatically recognize obscene words and their derivatives – the data will be taken from the list of obscene words and expressions of Roskomnadzor. However, the specialists of the online cinema do not plan to stop only on the definition of forbidden words.

Chekanov also said that the software created by the company will recognize smoking scenes in the frame, as well as the use of alcoholic beverages and scenes of a sexual nature. In theory, as the company said, in the foreseeable future, the software will be able to detect even scenes of violence, but this option, apparently, will be implemented later. The Megogo robot program will be built on a neural network, which in turn is based on a special speech recognition tool Vock (an open source solution) and the latest system for visual recognition of objects in the frame. True, the online cinema does not plan to rely solely on automation.

For example, the system might define the phrase “don’t freak out” as obscene, although this is clearly not the case. Accordingly, when filtering content, the video service will still involve human specialists who will view the actions of the neural network before delivering the final verdict. But in about a year, the technology will work in a completely autonomous mode without human intervention – the company is counting on this. Now the software will have to analyze more than 700 thousand films and series of TV series – obscene words will be “spiked”, while the rest of the content will receive the corresponding marking “18+”. And Viktor Chekanov also said that if the technology performs the task efficiently enough, then it will most likely be sold to other online cinemas, which also faced the need to censor content.

You may also like

Get the latest

Stay Informed: Get the Latest Updates and Insights

 

Most popular