How Our Technology Works

Video processing that unlocks context rapidly.

Video is inherently more complex to analyse than audio, but it’s essential for comprehensive broadcast monitoring. eMM’s video processing engine supports real-time analytics, worldwide multi-language coverage, instant alerts, story segmentation, and precise interest definitions, all derived from rich textual representations of both audio and visual content. 

• Key-Frame Detection:Identify the most significant and informative frames within a broadcast stream.
• Shot-Cut Detection: Automatically recognize scene transitions to support accurate story segmentation.
• Text Detection:Locate and extract on-screen text such as chyrons, banners, and lower thirds.
• Optical Character Recognition (OCR):Convert detected on-screen text into structured, searchable data for deeper analysis and reporting.

Keyframe Identification

Reduce volume, keep meaning.

Keyframe identification captures and stores a complete image from a broadcast stream, significantly reducing the volume of video data that needs to be processed and stored, without losing essential visual context or analytical accuracy.

Shotcut Detection

Understand where scenes change.

Broadcast content rarely stays on a single camera angle. Our Shot-cut detection engine automatically identifies abrupt transitions within a stream, pinpointing exactly where one shot ends and another begins, improving the precision of downstream analysis and reporting.

Text detection and recognition

Turn on-screen text into a searchable signal.

eMM captures audio and video from a vast array of sources, processing them through specialised components to extract every layer of information. By converting on-screen visual cues into text-based outputs, we enable high-scale use cases, including rapid information retrieval, topic tracking, automated summarization, and concept linkage. This structured data allows you to categorize, visualize, and query broadcast content with unprecedented scale and precision.

Format Converter

Make media usable across workflows.

For media to be truly actionable, it must be accessible. Our platform encodes and compresses content into high-performance, streamable formats, then intelligently segments it to support user-defined downloads. Whether you need a quick preview or a full-resolution export, our converter ensures consistent quality and seamless integration into your reporting tools.

Subtitle Reader

Don’t miss what’s already captioned.

When broadcasts provide subtitles or closed captions, eMM automatically extracts them and converts them into time-stamped, searchable text. We also enrich this layer with additional metadata, such as electronic programme guide (EPG) information to provide deeper context and more complete records.

Text Insert Reader

Capture lower-thirds and speaker IDs automatically.

Many programmes display text under speakers, such as names and roles. eMM’s text insert reader detects these visual text blocks and uses OCR to convert them into time-stamped text, adding crucial context that helps clarify who said what. 

Book a demo

Ready to see video processing on your channels?

Schedule a demo and we’ll show how eMM extracts searchable, time-stamped context from broadcast video, including keyframes, shot changes, subtitles, closed captions and on-screen text, so you can analyse coverage faster and with greater confidence.

Schedule a demo

Find more information about the other technologies used by eMM:

> Audio Processing

> Real Time

> Semantic Analysis