Dynamic Quantization Step Adjustment for Optical Black Noise
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image capturing devices face challenges in accurately estimating and suppressing noise in digital image data due to varying noise levels caused by image sensor characteristics and temperature conditions, especially with increased sensitivity, leading to decreased image quality.
Innovation Solution
An encoding apparatus and method that analyzes the optical black area of raw data to estimate noise levels, using different quantization steps based on noise amount indices to effectively suppress noise by adjusting quantization parameters during the encoding process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image capturing sensitivity is increased, then image quality is improved, but random noise becomes more noticeable
Solution Approach 1:
The patent changes the quantization step parameter based on the noise amount index value. When the noise amount is large (higher sensitivity), a larger quantization step is applied to suppress noise. When the noise amount is small (lower sensitivity), a smaller quantization step is used to maintain image quality. This dynamic parameter adjustment resolves the contradiction between image quality and noise visibility.
2Device complexity
If a fixed quantization table is used, then encoding is simplified, but noise suppression accuracy deteriorates
Solution Approach 1:
The patent introduces dynamic selection of quantization tables based on the noise amount index value. Instead of using a fixed quantization table, the system dynamically switches between different quantization tables (first quantization table for low noise, second quantization table for high noise) according to the actual noise conditions detected in the optical black area. This resolves the contradiction by making the encoding process adaptive to noise levels.
3Ease of operation
If noise amount estimation is based on image capturing conditions, then the process is simplified, but estimation accuracy deteriorates
Solution Approach 1:
The patent enables the system to self-measure the noise amount by analyzing the optical black area in the captured image data. Instead of relying on external or pre-set estimates based on capturing conditions, the system directly measures the actual noise present in the image by examining the optical black pixels. This self-measurement approach resolves the contradiction by providing accurate, condition-specific noise estimation without requiring complex external parameter tracking.
Data Source
AI summary
There is provided an encoding apparatus. An acquiring unit acquires an index value of a noise amount by analyzing at least a part of an optical black area that is included in raw data. A quantizing unit quantizes the raw data based on the index value. An encoding unit encodes the quantized raw data. If the index value is a first value, the quantizing unit quantizes the raw data with a first quantization step. If the index value is a second value corresponding to a larger noise amount than a noise amount in a case where the index value is the first value, the quantizing unit quantizes the raw data with a second quantization step that is larger than the first quantization step.