What is the Deep Media Analyzer?

Deep Media Analyzer provides best-in-class (best-of-breed) AI algorithms to enrich all your media assets with valuable metadata through a perfectly integrable, scalable, and secure solution. It automates the process of identifying, categorizing, and tagging content by recognizing various elements like faces, objects, actions, and more. It uses machine learning and computer vision and is the key component of our Composite AI platform and the foundation of many workflows.

What is the Deep Media Analyzer, and what can I do with it?

• It processes visual, audio, and metadata to extract insights.
• Recognizes and tags faces, objects, actions, text, and speech.
• Provides automatic keyword suggestions or summaries for better content searchability.
• Can integrate into workflows via APIs for seamless automation.

The Deep Media Analyzer incorporates several modules:

• Face Recognition:
  • Detects and identifies public figures in categories like politics, sports, and entertainment, leveraging a pre-trained dataset of over 20,000 personalities.

• Face Attributes:
  
  • Face Attributes recognizes emotions, ethnicity, gender or facial characteristics such as "beard", "eyes closed" or "glasses" of all persons appearing in pictures or videos

• Object & Scene Recognition:
  • Automatically detects and labels objects and scenes in videos or images. With over 1,500 object classes, it helps you categorize and archive visual data efficiently.

• Lower Third Recognition:
  • Reads on-screen text inserts (such as names) and associates them with corresponding faces in the video, helping in creating datasets of personalities.

• Speech Recognition:
  • Converts spoken language into text (speech-to-text), detects the spoken language, and can even recognize custom entities from a dictionary. It also supports transcription and translation.

• Personal Data Anonymization
  
  • Blurs faces and license plates in images and videos. It is a privacy protection tool that helps to comply with data protection regulations.

• Landmark Recognition:
  • Identifies well-known landmarks and architectural structures in visual data, helping to classify content based on geographic and cultural contexts.

• QR Code Detection:
  • Detects and decodes QR codes as well as EAN-13 codes (product barcodes) from videos and images, making it useful for media that includes scannable codes.
    

• Advanced Diversity Analysis:
  • Analyzes the percentage of gender and age representation in visual content, which can be useful for monitoring the diversity of characters in media productions.

• Subtitle Detection:
  • Detects burned-in subtitles in videos, identifies their position, language, and text content for further analysis or extraction.

• Text Recognition:
  
  • Find and extract characters or words in a media file

• Diversity Analysis (Legacy):
  • Diversity Analysis offers the possibility to determine the percentage of gender occurrence in images or videos.  This module was be replaced by the Advanced Diversity Analysis and will soon be deactivated.

Here are some of the key functionalities:

1. Content Recognition:

   • Deep Media Analyzer can identify and tag a wide variety of visual elements, such as objects, faces, landmarks, and even logos in your media. It automatically labels these elements, making it easier to search and categorize your content.

2. Speech-to-Text:

   • This feature allows the system to convert spoken words into text, creating transcripts that can be used for subtitle generation or further analysis. This functionality is particularly useful for broadcasts or interviews where dialogue needs to be transcribed quickly and accurately.

3. Contextual Tagging:

   • The AI can analyze the context of a scene and generate relevant keywords based on the visual and spoken content. For example, if a scene contains food-related objects, words like “fruit” or “salad” might be tagged. This can be used to optimize content for advertising or enhance recommendation systems.

4. Diversity and Equality Insights:

   • Deep Media Analyzer also offers insights into the gender and age representation in your media. It can break down how often certain groups appear on screen and how much they speak, making it a valuable tool for assessing diversity in your content.

5. Metadata Generation:

   • It automatically generates metadata for your media assets, allowing for better organization, searchability, and classification of large media libraries.

Use Cases:

• Broadcast Media: Automatically tag and transcribe live events, making content easily searchable and reducing the time needed for manual reviews.
• Advertising: Use contextual tagging to place targeted ads based on scene content, ensuring better audience engagement and maximizing ad relevance.
• Content Management: For industries like film production or news, Deep Media Analyzer helps in managing large archives, enabling quick searches based on visual or speech-based criteria.