Motion Picture Encoding Device Noise-Based Intra-Frame Coding Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing motion picture encoding devices perform excessive calculations for intra-prediction processing and fail to consider human visual characteristics, leading to reduced image quality and quantization error propagation due to the selection of suboptimal coding modes.
Innovation Solution
A motion picture encoding device that selectively chooses between intra-frame and inter-frame coding based on the presence of visually noticeable noise, using error estimation and flatness detection to determine the appropriate coding mode, thereby reducing wasteful calculations and preventing quantization error propagation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If intra-prediction processing is performed for all block images, then coding mode determination is thorough, but calculation complexity increases excessively
Solution Approach 1:
The patent performs preliminary detection of visually noticeable noise in block images before committing to full intra-prediction processing. By using error estimation and flatness detection as preliminary checks, the system identifies cases where intra-frame coding is necessary without performing all possible calculations, thus avoiding excessive computational complexity while maintaining accurate coding mode determination.
Solution Approach 2:
The patent applies partial action by performing only the necessary calculations based on detected noise characteristics. Instead of exhaustively evaluating all prediction modes, the system performs error estimation and flatness detection selectively, and only performs full intra-prediction processing when visually noticeable noise is detected, thereby reducing unnecessary calculations while maintaining sufficient accuracy.
2Productivity
If inter-frame coding is used for all block images, then calculation efficiency is improved, but quantization error propagation occurs reducing image quality
Solution Approach 1:
The patent implements feedback by continuously monitoring the decoded image for visually noticeable noise and using this information to adjust the coding mode selection. The error estimation and flatness detection provide feedback that triggers intra-frame coding when noise is detected, preventing quantization error propagation while maintaining inter-frame coding efficiency for clean images.
Solution Approach 2:
The patent applies preliminary anti-action by detecting visually noticeable noise before it can cause significant image quality degradation. By using error estimation and flatness detection to identify problematic block images in advance, the system preemptively switches to intra-frame coding to prevent quantization error propagation, rather than reacting after quality degradation has occurred.
3Ease of operation
If coding mode selection is based solely on coded data amount, then selection is simple, but human visual characteristics are not considered leading to suboptimal quality
Solution Approach 1:
The patent applies local quality by analyzing specific characteristics of individual block images, particularly flatness and noise distribution, rather than using a uniform selection criterion for all images. The error estimation and flatness detection examine local image properties to determine coding mode selection, thereby considering human visual characteristics and preventing quantization error propagation in flat regions while maintaining simplicity through targeted analysis.
Data Source
AI summary
The minimum Sum of Absolute Differences obtained by a motion vector search roughly judges the magnitude of quantization error by whether or not exceeding a predetermined threshold value. When the quantization error is lower, whether or not visually noticeable noise exists in some of the pixels of the current macroblock is judged based on the amount of flatness and noise detected in each of the 4x4 pixel blocks of the current macroblock partitioned into 16 sub-macroblocks. If there is visually noticeable noise, intra-frame coding is selected. When the quantization error is higher, whether or not visually noticeable noise exists in the current macroblock is judged while considering the magnitude of the motion vector. If there is visually noticeable noise, intra-frame coding is selected.