Image Quality Assessment via Adaptive Non-Overlapping Mean Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image quality assessment methods in marketing auditing are inefficient and resource-intensive, relying on manual reviews that are subjective and time-consuming, especially when dealing with large volumes of images from multiple locations, leading to inaccuracies and delays in providing feedback to auditors.
Innovation Solution
The implementation of adaptive non-overlapping mean estimation (ANME) techniques, which segment images into blocks, calculate mean pixel values, and convert them into smaller-sized images for feature extraction, followed by neural network classification, enabling objective and rapid assessment of image quality without a reference image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual review methods are used for image quality assessment, then subjective evaluation can be performed, but the process becomes time-consuming and resource-intensive
Solution Approach 1:
The patent segments images into non-overlapping blocks and applies mean estimation to each block, transforming the entire image into a smaller representation. This segmentation approach enables faster processing while maintaining quality assessment accuracy by focusing on representative block characteristics rather than processing every pixel manually.
Solution Approach 2:
The patent replaces manual mechanical review processes with an automated neural network system. The system automatically performs image quality assessment by processing segmented image blocks through trained neural networks, eliminating the need for time-consuming manual inspection while maintaining or improving assessment accuracy.
2Measurement precision
If manual image quality assessment is performed, then detailed evaluation can be conducted, but resource expenditure increases significantly
Solution Approach 1:
The patent extracts essential quality features from images by calculating mean values of segmented blocks and processing only these extracted features through neural networks. This extraction approach maintains detailed quality assessment capability while significantly reducing computational resources compared to processing entire high-resolution images manually or through resource-intensive algorithms.
Solution Approach 2:
The patent transforms images from their original high-dimensional pixel representation into lower-dimensional mean value representations of segmented blocks. This parameter transformation reduces the computational complexity and resource requirements while preserving the essential quality information needed for accurate assessment.
3Measurement precision
If traditional image processing methods are used, then comprehensive analysis can be performed, but processing speed decreases
Solution Approach 1:
By dividing images into non-overlapping blocks and processing each block independently through mean estimation, the patent enables parallel processing of multiple image regions. This segmentation strategy maintains comprehensive analysis capability while dramatically improving processing speed compared to traditional methods that process entire images sequentially.
Solution Approach 2:
The patent applies mean estimation to representative blocks rather than processing every pixel in detail. This partial action approach focuses computational effort on key representative regions, achieving sufficient assessment accuracy while processing images much faster than methods that attempt to analyze all pixels comprehensively.
4Measurement precision
If feedback is provided after auditing completion, then thorough evaluation can be given, but real-time feedback is delayed
Solution Approach 1:
The patent performs image quality assessment during the auditing process itself using automated neural network processing of segmented image blocks, rather than waiting until auditing completion. This preliminary action enables real-time feedback to be provided to auditors while they are still on-site, maintaining thorough evaluation capability while eliminating delays associated with post-auditing centralized processing.
Data Source
AI summary
Methods, apparatus, systems and articles of manufacture (e.g., physical storage media) to assess image quality using adaptive non-overlapping mean estimation are disclosed. Example apparatus disclosed herein include a machine learning system to be trained to classify image quality. Disclosed example apparatus also include a feature extractor to apply a blur filter to a first image to determine a blurred image, determine a blur feature value for the first image, the blur feature value to represent an amount the first image differs from the blurred image, and apply a vector of feature values associated with the first image to the machine learning system, the vector of feature values including the blur feature value. Disclosed example apparatus further include an image classifier to classify image quality associated with the first image based on an output of the machine learning system responsive to the vector of feature values associated with the first image.