Region-Based Cadence Detection for Video Signal Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing cadence detection methods for sequences of images are inefficient, particularly in identifying film and video formats, leading to unnecessary application of complex deinterlacing algorithms and increased data transmission, which can result in image artifacts and reduced quality.
Innovation Solution
A region-based cadence detection method that compares pixels in current fields to previous fields, assigns motion phase values, and segments fields into regions based on block motion phase values, reducing data transmission by sending only region motion phases and indicators, rather than per-pixel or per-block data, and utilizing previous segmentation and motion phase history for stability and reduced resource usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If per-pixel or per-block cadence detection is performed, then detection precision is improved, but data transmission volume increases and processing complexity increases
Solution Approach 1:
The patent segments the image field into multiple blocks and performs cadence detection at the block level rather than per-pixel. Each block generates a motion phase value representing the motion state of all pixels within that block. This segmentation approach maintains detection precision by capturing motion characteristics at a sufficient granularity while dramatically reducing the volume of data that needs to be transmitted and processed, as entire blocks are represented by single motion phase values.
2Reliability
If complex deinterlacing algorithms are applied to all fields, then image quality is improved, but processing time increases and unnecessary processing occurs
Solution Approach 1:
The patent applies deinterlacing processing selectively based on local motion characteristics. By calculating motion phase values for each block and comparing them across fields, the system identifies regions where motion occurs and applies deinterlacing only to those specific blocks. Blocks with no motion detected between fields are processed more simply. This local quality approach ensures image quality is maintained in regions where it matters while avoiding unnecessary processing in static regions, thereby reducing overall processing time.
Solution Approach 2:
Instead of applying full deinterlacing algorithms to all fields unconditionally, the patent performs partial processing by first conducting a rapid motion analysis at the block level. Only blocks exhibiting motion characteristics undergo the more computationally intensive deinterlacing operation. This partial action strategy ensures that processing resources are concentrated where actually needed, eliminating wasted computation on static portions of the image while maintaining quality where motion requires it.
3Measurement precision
If motion analysis is performed on all pixels, then detection accuracy is improved, but computational resources increase
Solution Approach 1:
The patent merges the motion information of all pixels within a block into a single motion phase value by comparing the current field block with the corresponding previous field block. This merging process involves calculating the difference between corresponding pixel positions and aggregating this information into one representative value for the entire block. By combining multiple pixel measurements into a single metric, the system maintains motion detection accuracy at the block level while reducing computational energy requirements, as processing one aggregated value consumes far less energy than processing each individual pixel separately.
Data Source
AI summary
In a method for detecting the cadence of a sequence of images, each pixel in each current field in the sequence of images is compared to at least one pixel in at least one previous field. A pixel motion phase value is assigned to each pixel in the field as a function of the result of that comparison. For each block of pixels in the current field, a block motion phase value is determined from the motion phase values of the pixels in the block. The current field is segmented into at least one region, with each region comprising a whole number of blocks, as a function of at least the determined block motion phase values. A region motion phase value is assigned to each region, based on the block motion phase values for the blocks in the region.