Connections

AWS Elemental Cloud Services Aim to Maximize Value of Video with AI

NEW YORK — Four of the most recent cloud services that Amazon Web Services (AWS) introduced are helping media and entertainment (M&E) companies maximize the value of video with artificial intelligence (AI) and machine learning (ML), according to Evan Statton, senior solutions architect at AWS Elemental.

He was referring to Amazon Comprehend, Amazon Rekognition Video, Amazon Transcribe and Amazon Translate.

While introducing those services, the company said at the time that they allow developers to build applications that emulate human-like cognition: Comprehend for understanding relationships and finding insights within text, Transcribe for converting speech to text, Translate for translating text between languages, and Amazon Rekognition, a deep-learning powered video analysis service that Amazon said “tracks people, detects activities, and recognizes objects, celebrities, and inappropriate content.”

Amazon Comprehend enables you to “discover insights about your content,” Statton said at a media conference May 7, noting that “usually it’s going to be text-based.” While using Comprehend, users provide a list of text data and the service will” find interesting terms in the text that it thinks go together and have some relevance,” he said.
Rekognition handles both video and still image analysis, he pointed out, noting it’ll “tell you what’s in a frame, for example.”

Comprehend, Rekognition and Transcribe can all be combined together, he said.

For example, translating speech to text and combining it with Comprehend may tell the user “something that somebody is saying, and you can use that as metadata,” he said.
Common M&E use cases for Transcribe include automated closed caption and subtitle generation, audio content indexing and compliance/objectionable content detection, he told the conference.

“Once you have [the text] in English, then you can translate it to any number of languages that AWS supports” with Amazon Translate, he said.

The AWS cloud services apply “to other markets besides just the traditional M&E space,” he went on to say. As examples, Rekognition Video can be used in home security for safety and detection and in the government security sector for surveilling people, he noted. “If you have a squirrel that keeps eating your bread, then you can definitely find that squirrel and get rid of it,” he joked.

Meanwhile, the education and enterprise markets can use the services for global reach, including Transcribe for audio and Translate for different languages, according to Statton. “We’re only limited by our imaginations here,” he concluded his presentation saying, adding: “You could really take these building blocks … and piece them together any way you’d like.”

At the same conference, Tom Sauer, chief development officer at ZoneTV, predicted that traditional pay TV operators will “play an important role for a long time” despite the growing popularity of over-the-top (OTT) streaming services. “It’s a great time to be in the TV business,” he said.

While introducing a new linear pay TV service making use of AI in a unique way that Zone TV predicted could serve as a template to help operators overcome the challenges impacting the pay TV industry today, CEO Jeff Weber said last year: “You have these two worlds that have completely different economics” in the pay TV and OTT space and, “from a consumer perspective, it makes no sense at all.”

As part of ZoneTV Dynamic Channels, ZoneTV’s licensed, digital-first content is being curated into specialized channels for delivery to pay TV subscribers, and combines linear, on-demand and customized content selections, to create “a unique and personalized experience for the consumer,” ZoneTV and partner Ooyala said in a news release in July 2017.

ZoneTV uses Ooyala’s content management Flex platform for end-to-end video workflow for both new specialized channels and video on demand assets, the companies said. Flex runs on Microsoft Azure Media Services and can transform assets into any format for distribution to any device with Ooyala’s video platform, handling quick transcoding, packaging and syndication, to enable quick delivery of content to audiences.

“We’ve got to create more dynamism within [the] old TV space” and “we believe [ZoneTV’s offering] is a first step in creating that next level of” dynamism, Sauer said May 7. The service is handled via OTT, so “there’s no need for new hardware,” he noted.