Video Encoder Quantization Using Human Visual Sensitivity Deadzones
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Encoders for high-resolution video content lack consideration of human visual properties, leading to inefficient data usage and suboptimal image quality during quantization.
Innovation Solution
An encoder is designed with a domain transform unit, a deadzone determination unit based on a Human Visual System (HVS) table, and a quantization unit that sets a deadzone for frequency data, prioritizing human visual sensitivity to reduce unnecessary data transmission and enhance image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional quantization is used without considering human visual properties, then the encoding process is simple, but the image quality and data efficiency are suboptimal
Solution Approach 1:
The patent applies parameter changes by modifying the quantization step size dynamically based on frequency characteristics and human visual sensitivity. Different quantization parameters are selected for different frequency bands, allowing finer quantization for visually important frequencies and coarser quantization for less important frequencies, thereby improving image quality while maintaining encoding efficiency
Solution Approach 2:
The patent implements local quality by applying different quantization strategies to different frequency components. High-frequency components that are less sensitive to human vision receive coarser quantization, while low-frequency components that are more visually important receive finer quantization. This localized approach optimizes overall image quality by allocating precision where it matters most
2Quantity of substance
If all frequency data is transmitted with equal precision, then the encoding is straightforward, but the data transmission volume is huge
Solution Approach 1:
The patent changes the quantization parameter based on frequency characteristics and human visual sensitivity. For frequencies that are less sensitive to human perception, larger quantization steps are used, reducing the precision requirement and thereby decreasing the data transmission volume. This selective parameter adjustment maintains acceptable image quality while significantly reducing bandwidth requirements
Solution Approach 2:
The patent extracts and identifies the essential visual information by analyzing human visual sensitivity characteristics. It separates frequency components into categories based on their visual importance, transmitting only the necessary precision for each category. This extraction of essential information allows for reduced data transmission volume without sacrificing perceived image quality
3Productivity
If coarse quantization is applied to reduce data amount, then transmission efficiency improves, but image quality deteriorates
Solution Approach 1:
The patent applies local quality by differentiating quantization precision across frequency bands. Coarse quantization is applied only to frequency components that are less sensitive to human vision, while fine quantization is maintained for visually important components. This localized differentiation achieves good transmission efficiency without compromising overall image quality, as the coarse quantization is applied where it is least noticeable
4Quantity of substance
If traditional encoders process high-resolution video, then the resolution is maintained, but the data usage is huge
Solution Approach 1:
The patent changes quantization parameters dynamically based on the frequency characteristics of the high-resolution video content and human visual sensitivity. By adapting the quantization step size to the actual content and perceptual importance, the encoder achieves efficient compression of high-resolution video without requiring excessive data transmission, thereby improving encoding efficiency while reducing data usage
Data Source
AI summary
An encoder that performs quantization based on a deadzone and a video processing system including the encoder are provided. The encoder includes: a domain transform unit configured to transform first image data of a spatial domain into second image data of a frequency domain including first to N-th pieces of frequency data (wherein N is an integer equal to or greater than 1); a deadzone determination unit including a human visual system (HVS) table including human visual frequency sensitivities and configured to determine, based on the HVS table, a deadzone regarding each of the first to N-th pieces of frequency data; and a quantization unit configured to perform, based on the deadzone, quantization regarding the second image data.