In Broadcast - July 2021 - 65

www.inbroadcast.com | Vol: 11 - Issue 7 | July 2021
65
InMedia
The benefits of using AI
for content analysis
By Rick Capstraw, Chief Revenue Officer...
he use cases for Artificial
Intelligence (AI) in the broadcast
industry are growing rapidly with
every month. Its initial deployments
have largely been within the field
of
T
automation,
however, and it
is
currently here that we find it making its
deepest impact.
AI is deeply embedded within a tool
called Deep Analysis that we bundle
with OWNZONES Connect, our content
servicing and supply chain platform in
the cloud. Using it, broadcasters are
able to not only automate critical parts
in the broadcast workflow, they are
able to dramatically speed them up.
For example, it can be used to scan
content upon ingest to understand its
layout and formatting, what we call
the FrameDNA. Understanding this
initial layout and format is critical to
compliance and localisation
editing,
a necessary but painful task. Without
this FrameDNA, operators need to
watch the full length of any ingested
content in order to identify scenes
or clips that need to be
adjusted, such as removing
commercial black sequences.
Using Deep Analysis we
can greatly accelerate the
manual process. Types of
scenes that can be detected
include slates, bars, blacks,
program
change
segments,
detection,
scene
texted
scenes, essences, and more.
When the AI detects these scenes,
which it does in much faster than
real-time, they are flagged up with
a marker on a timeline view. This not
only greatly increases the efficiency of
the compliance process, but that single
scan can be output as a file allowing
broadcasters to save the resulting data
for any future localisation, compliance
editing, delivery, and archival needs.
Text collision detection is another
process where users can gain hugely
from this deployment of AI. One of the
issues that can often occur in QC is that
the Timed Text will overlap with text
graphics in a scene, or the Timed Text
will extend into a further scene. This
can lead to rejections upon delivery to
a platform.
Deep Analysis Timed Text QC is an
AI-powered tool that analyzes timed
text positions against onscreen vector
images and displays results on a
markers bar so the user can see where
the timed text and the graphic text are
colliding. It also highlights the duration
of the timed text against the duration
of the scene and flags whether
subtitles are extending beyond a scene
change by analyzing the timed text's
begin and end points, and duration,
and compares this against scene
change data.
We estimate that under typical
circumstances Deep Analysis
works
more than 5x faster than real-time.
AI not only saves time though, it
eliminates errors. In developing
the tool in response to a particular
customer's needs, it detected 60%
more issues than had been picked up
in the manual QC process.
Increased efficiency and increased
accuracy? It is no wonder that AI
is providing increasingly popular
solutions that broadcasters are
seeking out.
https://ownzones.com
Open your compliance recording to your
entire organization for Automation
and Integrations
Use your compliance for cross organization
automatic workflows and benefit from integrations
with your other systems...
In today's world, every two
different applications must be able
to communicate with each other.
Any system must be able to integrate
with third parties systems, and vice
versa, through advanced API, to allow
integrations and automatic workflows.
APIs are more than technical
" communication " for sharing
data. Well-designed APIs provide
organizations with a critical link to
data and services that enable rapid
innovation and open new workflows
and services. This is exactly what Actus
API was designed for.
Actus compliance recording and
intelligent monitoring platform is
an open system. Our rich, scalable
and authenticated REST API is an
integral part of it. We integrate with
any third-party system and allow
any third party to integrate with our
platform capabilities and data. Our API
allows automatic workflows for better
efficiency, faster operations, saving
human resources, avoid errors and
provides a new source of information.
It goes beyond the fact that all Actus
data is standard, open and can be
accessed and monitored with standard
tools.
For example, all the recorded media
can be accessed by any standard
video player, or the database is purely
open. A closed application that is
using proprietary data and is lacking
the ability to communicate with other
applications is not viable.
Here are a few examples of Actus
platform use cases that highlights the
importance and added values of the
above.
Automatic clipping - Actus provides a
complete automatic clipping workflow.
All the functionalities are achieved
through Actus API, integration to
playout systems metadata, advanced
clipping options and extracted
metadata, third parties' applications
such as external transcoding farms.
Such a workflow will speed up the
process, save human costs and will be
more accurate.
Automatic reports based on the
playout systems metadata such as
reports on the ads that have been
aired , discrepancy reports based on
the playout data and scheduling data.
Automatic scheduled clipping - no
need to clip specific content manually.
With Actus API and defined set of rules
the clips will be created and exported
to the right destination in a complete
automated workflow.
Actus player API allows any other
application
to
benefit
common
from
use
Actus
smart player and unique features.
Another
case is
integration with MAM systems. One
example is loading automatically
the MAM Ads files for the purpose of
Actus automatic ads detection. The
automatic ads detection allows to
track the aired ads on competitive
channels without the need to load
one Ad at a time, saving time and
expensive human resources.
https://actusdigital.com
http://www.inbroadcast.com https://ownzones.com https://ownzones.com https://ownzones.com https://www.ownzones.com https://actusdigital.com https://actusdigital.com https://www.actusdigital.com
In Broadcast - July 2021

Table of Contents for the Digital Edition of In Broadcast - July 2021
