Block Boundary Detection in Scaled Video Using DCT Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional video processing techniques fail to detect block boundaries accurately when blocks are not of a fixed or known size, particularly in scaled video data, leading to missed detections and false positives due to varying block sizes and compression artifacts.
Innovation Solution
A video processing system that includes a block boundary detector, array generator, DCT module, and post-processor to analyze luminance data, generate detection arrays, compute DCT coefficients, and identify the presence and location of block artifacts, using techniques like high-pass filtering, threshold quantization, and discrete cosine transformation to determine block sizes and reduce visibility of block noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional video processing techniques are used to detect block boundaries, then the detection process is simple, but the detection accuracy deteriorates when blocks are not of fixed or known size
Solution Approach 1:
The video frame is divided into multiple blocks of varying sizes, and each block is processed separately through the detection pipeline (high-pass filtering, threshold quantization, DCT transformation). This segmentation allows the system to handle different block sizes independently, improving detection accuracy without requiring a single complex algorithm for all sizes
Solution Approach 2:
The patent transforms the block detection problem from spatial domain analysis to frequency domain analysis using Discrete Cosine Transformation. By converting block boundary detection into frequency spectrum analysis, the system can effectively detect boundaries regardless of block size, as the periodicity patterns in the frequency domain reveal block structures that are not obvious in the spatial domain
2Measurement precision
If high-pass filtering and DCT transformation are applied to detect block boundaries, then detection accuracy improves, but processing time increases
Solution Approach 1:
High-pass filtering is applied as a preliminary step before DCT transformation to pre-process the luminance data and emphasize edge information. This preliminary action prepares the data in a way that makes subsequent DCT transformation more efficient and effective, reducing the computational burden of the main detection algorithm while maintaining high accuracy
Solution Approach 2:
The patent extracts only the essential frequency components through DCT transformation that are relevant for block boundary detection. By focusing on specific frequency coefficients that indicate block periodicity rather than processing the entire frequency spectrum, the system achieves accurate detection with reduced computational complexity and processing time
3Adaptability or versatility
If block boundaries are detected in scaled video data with varying block sizes, then adaptability improves, but false detections increase
Solution Approach 1:
The detection system dynamically adapts to different block sizes by analyzing the periodicity patterns in the DCT frequency spectrum. Rather than using fixed detection thresholds or assumptions about block dimensions, the system identifies the actual block size present in each video frame through frequency analysis, allowing it to accurately detect boundaries regardless of scaling or varying block sizes while maintaining high reliability
Solution Approach 2:
The system uses the periodicity information extracted from DCT coefficients as feedback to refine block boundary detection. By analyzing the frequency spectrum and identifying dominant periodic patterns, the system can adjust its detection strategy to match the actual block structure in the video data, reducing false detections while maintaining adaptability to different block sizes and scaling factors
Data Source
Figure 1A
Figure 1B
Figure 2A
AI summary
Systems and methods are provided for determining characteristics of video data. A frame of video data is obtained, where the frame is represented by pixel data. A value is assigned to an element of a detection array based on pixel data in a portion of the video frame corresponding to the element. A frequency transform of values of the detection array is determined, and a characteristic of the video data is extracted based on the output of the frequency transform.