JND Thresholds for Image Complexity Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing content delivery systems face challenges in efficiently reducing the size of high-quality images and videos for transmission without impacting their perceivable quality, as they often contain unnecessary information that increases storage and processing demands.
Innovation Solution
The system preprocesses content items by applying just-noticeable-difference (JND) thresholds to discrete cosine transform (DCT) sub-bands, setting values below perception thresholds to zero, and then inverting the transform to reduce complexity, using a combination of contrast sensitivity function, contrast masking, and luminance adaptation models to determine these thresholds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If high-quality images and videos are transmitted without preprocessing, then image quality is maintained, but file size and transmission bandwidth requirements increase
Solution Approach 1:
The patent extracts and removes unnoticeable information from images by applying JND thresholds to DCT sub-bands. Specifically, it identifies and eliminates frequency components below the just-noticeable-difference threshold, thereby reducing file size while preserving perceptible image quality.
Solution Approach 2:
The patent transforms image parameters by converting spatial domain pixel values to frequency domain DCT coefficients, then applies thresholding based on JND parameters. This parameter transformation enables selective removal of imperceptible frequency components while maintaining perceptible quality.
2Quantity of substance
If traditional encoding techniques are used, then images are compressed, but noticeable quality degradation occurs
Solution Approach 1:
The patent introduces JND threshold models as an intermediary between the original image and the compressed representation. These models serve as a mediator to identify which frequency components can be removed without affecting perceptible quality, enabling lossless compression from a perceptual standpoint.
Solution Approach 2:
The patent performs preliminary preprocessing of images by removing unnoticeable information before the main encoding process. This preliminary action reduces the amount of data that needs to be encoded, thereby improving overall compression efficiency without sacrificing perceptible quality.
3Loss of information
If all image information is retained for transmission, then complete image data is available, but storage and processing demands increase
Solution Approach 1:
The patent applies different quality standards to different frequency components of the image. High-frequency components above the JND threshold are preserved with high fidelity, while low-frequency components below the threshold are removed or coarsely quantized, creating a non-uniform quality distribution optimized for human perception.
Solution Approach 2:
The patent discards unnoticeable information (frequency components below JND thresholds) that cannot be recovered, while preserving noticeable information that can be recovered through the inverse DCT transform. This selective discarding reduces processing demands while maintaining essential image quality.
Data Source
AI summary
Systems and methods are described to process image-based content to reduce the complexity of information used to represent the content, without visible different to a viewing user. Individual images of the content items can be broken into image blocks, each of which can be converted via DCT algorithm to a DCT matrix. Values within the DCT matrix can be compared to just-noticeable-difference (JND) thresholds, below which information is unlikely to be perceptible to viewing users. For each value in the DCT matrix below a corresponding JND threshold, the value can be reduced to zero, thus reducing complexity of the image without perceptible difference to a viewing user and enabling more efficient use of computing resources to store and transmit the image. JND thresholds can be based on one or more models of human perception, including a contrast sensitivity function model, a luminance adjustment model, and a contrast masking model.