Display Image Error Detection Using Foreground-Background Dissimilarity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing error detection methods, such as CRC, are inadequate for detecting errors in display images after image processing, as they cannot effectively identify subtle changes that may affect user recognition of icons or images, and bit-wise error detection is not suitable for determining errors that impact user perception.
Innovation Solution
A circuit device and method that acquire image data, perform image processing, and calculate indices representing the degree of dissimilarity or matching between foreground and background images based on pixel values, allowing for error detection by comparing these indices with threshold values, enabling the detection of errors that affect image visibility and shape recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If bit-wise error detection such as CRC is used, then data transmission errors can be detected, but it cannot properly detect errors that affect user recognition of display images
Solution Approach 1:
The patent changes the detection parameter from bit-level CRC values to pixel-level image quality metrics. Instead of checking individual data bits, the system evaluates pixel values, color accuracy, and image characteristics to detect errors that actually affect display quality and user perception.
Solution Approach 2:
The patent replaces the mechanical bit-wise comparison system with an image processing-based detection system. It uses pixel value analysis, color space conversion, and image quality assessment algorithms to substitute the traditional CRC verification mechanism, enabling detection of visually significant errors.
2Reliability
If CRC error detection is performed on received image data, then data integrity can be verified, but error detection cannot be performed after image processing
Solution Approach 1:
The patent performs preliminary error detection by capturing a reference image before image processing operations. This reference image serves as a baseline for subsequent comparison, enabling error detection to be applied after various image processing steps such as scaling, rotation, or format conversion.
Solution Approach 2:
The patent implements a feedback mechanism where the detected image quality metrics are compared against threshold values. When errors are detected, the system can trigger corrective actions or alert the user, creating a closed-loop error detection system that adapts to different processing scenarios.
3Measurement precision
If minute errors such as one-bit errors are detected, then data accuracy is maintained, but the displayed image does not substantially differ and user recognition is not affected
Solution Approach 1:
The patent applies local quality assessment by evaluating different regions of the image with different importance weights. Critical areas such as text, icons, or focal points receive higher inspection weights, while less important background areas receive lower weights, allowing the system to focus detection resources on visually significant elements.
Solution Approach 2:
The patent uses partial action by selectively detecting only those errors that exceed a certain visual significance threshold. Instead of flagging every minor bit error, the system identifies and reports only those errors that are likely to affect user perception, reducing false alarms while maintaining detection effectiveness.
Data Source
AI summary
Provided is a circuit device that can appropriately detect an error in a display image obtained by performing image processing, an electronic apparatus, an error detection method, and the like. A circuit device includes an image acquisition circuit configured to acquire image data, and an index acquisition circuit configured to obtain an index for performing error detection on a display image based on the image data. The index acquisition circuit obtains an index that represents the degree of dissimilarity between a foreground image, which is an image of a given region of the display image, and a background image, of the display image, corresponding to a background of the foreground image. The index is obtained based on pixel values of the display image.