A Method for Detecting a Commercial in a Video Data Stream by Evaluating Descriptor Information

Simple SummaryContent extracted from patent full text and abstract with AI.

This patent describes a method for automatically detecting commercials in a video data stream using multiple visual descriptors such as static logos, separating black frames, and hard-cut frequency. By analyzing video frames for these features and combining the results through an entropy-based decision tree (such as the ID3 algorithm), the system can accurately and efficiently identify segments that are likely commercials, suitable for real-time applications.

Use CasesContent extracted from patent full text and abstract with AI.

• Automatic commercial skipping or muting in set-top boxes or smart TVs.
• Video content analysis and tagging for video-on-demand platforms to enable ad-free viewing.
• Broadcast monitoring systems to log and verify commercial airtime.
• Assisting content aggregators in automatically segmenting and categorizing video streams.
• Enhancing user experience in personal video recorders (PVRs) by auto-marking commercial breaks.
• Supporting compliance and ad-insertion auditing for broadcasters and advertisers.

BenefitsContent extracted from patent full text and abstract with AI.

• Highly accurate commercial detection (up to 93% correct classification) with low false positives for other genres.
• Fast and computationally efficient, enabling real-time processing of live or recorded video streams.
• Reduces dependence on complex and expensive logo recognition algorithms.
• Minimizes false acceptance of fades or dark frames that aren't actual commercial breaks.
• Adaptable to various video sources and resolutions, including broadcast TV and internet streaming.
• Facilitates user-friendly video navigation and content personalization (e.g., genre-based or ad-free viewing).
• Automates tasks that traditionally required manual video review.

Technical Classifications (CPCs)

Inventors & Applicants

Patent Abstract

Systems, methods, and devices for detecting a commercial in a video data stream by evaluating descriptor information provided from a plurality of visual descriptors indicative of com- mercial or non-commercial features. The descriptor information provided from the plurality of descriptors may be evaluated in an entropy based decision process.