Sport Video Scoreboard OCR for Accurate Data Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for extracting sport-related information from digital video frames, such as scores and game times, are inefficient and prone to errors due to the recognition of unrelated text elements, leading to computational inefficiencies and inaccuracies.
Innovation Solution
A computing system analyzes multiple video frames to identify a scoreboard region based on static image elements, detects sub-regions containing text, and applies optical character recognition (OCR) to generate accurate sport data, focusing on relevant frame regions by leveraging the unchanging nature of scoreboard elements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the system analyzes all text elements in video frames to extract sport-related information, then comprehensive data extraction is achieved, but computational load increases and accuracy decreases due to recognition of unrelated text elements
Solution Approach 1:
The patent segments the video frame analysis by identifying and isolating the scoreboard region as a distinct area of interest. Instead of analyzing all text elements throughout the entire video frame, the system divides the frame into relevant (scoreboard) and irrelevant (other regions) segments, applying text recognition only to the scoreboard segment. This segmentation reduces computational load while maintaining extraction accuracy.
Solution Approach 2:
The patent applies local quality by treating the scoreboard region with specialized analysis methods different from the rest of the frame. The scoreboard region receives focused attention with dedicated text detection and recognition algorithms, while other regions are either analyzed with simpler methods or excluded entirely. This localized approach optimizes resources by applying high-precision processing only where needed.
2Speed
If the system processes every frame to ensure real-time data extraction, then responsiveness is improved, but computational resources are overwhelmed leading to inefficiencies
Solution Approach 1:
The patent extracts only the essential scoreboard region from each video frame, removing unnecessary processing of other frame contents. By taking out and isolating the scoreboard area using image processing techniques, the system processes only the relevant portion of each frame, maintaining real-time responsiveness while significantly improving computational efficiency.
Solution Approach 2:
The patent performs preliminary actions by pre-identifying the scoreboard region in each frame before conducting text recognition. The system first locates the scoreboard using image processing (such as detecting rectangular boundaries or specific color patterns), then proceeds with OCR only on the identified scoreboard area. This preliminary region identification streamlines the overall processing pipeline.
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
A computing system determines accuracy of sport-related information extracted from a time sequence of digital video frames that represent a sport event, the extracted sport- related information including an attribute that changes over the time sequence, The computing system (a) detects, based on the extracted sport-related information, a pattern of change of the attribute over the time sequence and (b) makes a determination of whether the detected pattern is an expected pattern of change associated with the sport event, lf the determination is that the detected pattern is the expected pattern, then, responsive to making the determination, the computing system takes a first action that corresponds to the sport- related information being accurate. Whereas, if the determination is that the detected pattern is not the expected pattern, then, responsive to making the determination, the computing system takes a second action that corresponds to the sport-related information being inaccurate.