Video Text Detection via Frame Pre-Filtering

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Viewers have limited options for consuming text in video content, such as television programs and commercials, with no efficient way to extract and utilize the information presented.

Innovation Solution

A text detection system that identifies and interprets overlaid text in video content, using techniques like optical character recognition, to extract information like URLs, user accounts, and phone numbers, and provides it to media content devices for display or further processing.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If text detection is performed on all video frames, then text information can be extracted, but computational cost and processing time increase significantly

Engineering Contradiction:
Improvetext information extractionVSAvoidcomputational efficiency
Core Design Contradiction:
Loss of informationVSProductivity

Solution Approach 1:

The system performs preliminary analysis on video frames to identify those likely to contain text before applying full text detection. This pre-filtering step examines frame characteristics such as motion patterns, region of interest detection, and visual features to predict text presence, thereby avoiding computationally expensive processing on frames without text while ensuring text-containing frames are captured for analysis

Inventive Principle:
Principle #10Preliminary action

2Loss of information

If comprehensive text detection is applied to all video content, then all text information is captured, but processing time and computational resources are wasted on frames without text

Engineering Contradiction:
Improvetext information completenessVSAvoidprocessing time
Core Design Contradiction:
Loss of informationVSLoss of time

Solution Approach 1:

The system performs preliminary analysis on video frames to identify those likely to contain text before applying full text detection. This pre-filtering step examines frame characteristics such as motion patterns, region of interest detection, and visual features to predict text presence, thereby avoiding computationally expensive processing on frames without text while ensuring text-containing frames are captured for analysis

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system applies text detection selectively to only those frames predicted to contain text, rather than performing comprehensive detection on all frames. This partial action approach processes a subset of frames (those with high probability of containing text) while skipping frames unlikely to have text, thus reducing overall processing time and computational resource consumption while maintaining text information completeness

Inventive Principle:
Principle #16Partial or excessive action

3Loss of information

If text detection system processes every frame, then no text is missed, but computational expenses increase

Engineering Contradiction:
Improvetext detection accuracyVSAvoidcomputational cost
Core Design Contradiction:
Loss of informationVSProductivity

Solution Approach 1:

The system performs preliminary analysis on video frames to identify those likely to contain text before applying full text detection. This pre-filtering step examines frame characteristics such as motion patterns, region of interest detection, and visual features to predict text presence, thereby avoiding computationally expensive processing on frames without text while ensuring text-containing frames are captured for analysis

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system applies text detection selectively to only those frames predicted to contain text, rather than performing comprehensive detection on all frames. This partial action approach processes a subset of frames (those with high probability of containing text) while skipping frames unlikely to have text, thus reducing overall processing time and computational resource consumption while maintaining text information completeness

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3149658B1Text detection in video
Publication Date: 2020.07.29 GRACENOTE INC
  • EP3149658B1 patent drawingFigure 1
  • EP3149658B1 patent drawingFigure 2
  • EP3149658B1 patent drawingFigure 3A

AI summary

Techniques of detecting text in video are disclosed. In some embodiments, a portion of video content can be identified as having text. Text within the identified portion of the video content can be identified. A category for the identified text can be determined. In some embodiments, a determination is made as to whether the video content satisfies at least one predetermined condition, and the portion of video content is identified as having text in response to a determination that the video content satisfies the predetermined condition(s). In some embodiments, the predetermined condition(s) comprises at least one of a minimum level of clarity, a minimum level of contrast, and a minimum level of content stability across multiple frames. In some embodiments, additional information corresponding to the video content is determined based on the identified text and the determined category.