Telop Classification by Image and Semantic Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional techniques for collecting and displaying telops in television programs are inefficient in handling character strings beyond titles, resulting in a large number of unnecessary telops being included in user lists, which hinders easy recognition and reuse of video image content.
Innovation Solution
A telop collecting apparatus and method that performs image analysis on telop regions for coordinates, size, and color scheme, followed by character recognition and semantic analysis to classify telops into categories, allowing for the accumulation of telop information as text data organized by category.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If conventional telop collection techniques are used, then telops are collected and displayed in a list, but a large number of unnecessary telops (non-titles) are included in the list, reducing usability
Solution Approach 1:
The patent segments telops into different categories (title telops vs. non-title telops) based on their appearance timing relative to scene changes. By dividing the telop collection into meaningful segments, the system can filter and organize telops to include only relevant title information in the user list, thereby improving usability while maintaining an appropriate quantity of useful telops.
2Loss of information
If all telops are collected without differentiation, then comprehensive telop data is obtained, but the list becomes difficult to use for recognizing video content
Solution Approach 1:
The patent performs preliminary classification of telops by analyzing their appearance timing relative to scene changes before generating the final list. This preliminary action of categorizing telops as title or non-title allows the system to preserve complete telop data for analysis while selectively presenting only useful title telops in the user-facing list, thus maintaining information completeness while improving ease of content recognition.
Data Source
AI summary
According to one embodiment, a telop display system includes an extracting module which extracts a telop region identified as an image of a telop from video image information of a television program, an image analyzing module which performs image analysis related to coordinates, a size, and a color scheme of the telop region extracted by the extracting module, a semantic analyzing module which performs text analysis related to a word class and a meaning of the obtained character string, and a classifying module which classifies the telop on the basis of an analysis result of at least one of the image analysis and the text analysis to accumulate character strings of the telops as items of text information classified in units of categories.