Image Processing Apparatus Streak Noise Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing systems fail to accurately distinguish between streak-like noise and document edges, leading to incorrect removal of document edges during image cropping, resulting in undesirable output images with background noise included.
Innovation Solution
An image processing apparatus that employs two distinct methods for detecting and removing streaks from images: one using difference images in both longitudinal and transverse directions, and another using a histogram of edge pixel counts, ensuring accurate separation of document edges from noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single method for detecting streaks is used, then the processing is simple, but the accuracy of distinguishing streaks from document edges deteriorates
Solution Approach 1:
The patent applies segmentation by dividing the streak detection process into two distinct methods: (1) difference image method that compares adjacent lines to detect streaks, and (2) histogram method that analyzes pixel value distributions. Each method segments different aspects of streak characteristics, and their results are combined to achieve accurate detection while maintaining reasonable processing complexity through modular implementation.
Solution Approach 2:
The patent utilizes parameter changes by detecting streaks through multiple different detection parameters and methods. The difference image method detects changes in pixel values between adjacent lines, while the histogram method detects changes in pixel value distributions. By changing detection parameters and methods, the system achieves higher accuracy in distinguishing streaks from document edges.
2Object-generated harmful factors
If streak detection is performed aggressively, then streak removal is improved, but document edge preservation deteriorates
Solution Approach 1:
The patent applies feedback by using multiple detection methods that provide complementary information about streak presence. The difference image method provides feedback on local pixel value changes, while the histogram method provides feedback on overall pixel value distributions. This multi-feedback mechanism enables the system to distinguish between actual streaks and document edges more reliably, preventing erroneous removal of document edges while effectively removing streaks.
Solution Approach 2:
The patent uses partial action by applying different detection thresholds and methods to different regions of the image. The difference image method is applied to detect local streak variations, while the histogram method is applied to detect overall image characteristics. This partial application of different detection strategies allows effective streak removal while preserving document edges that may be mistaken for streaks in individual methods.
3Ease of operation
If document range detection is performed on images with streaks, then processing is straightforward, but detection accuracy deteriorates
Solution Approach 1:
The patent applies preliminary action by performing streak removal processing before document range detection. The controller first detects and removes streaks from the scanned image using the difference image method and histogram method, then performs document range detection on the cleaned image. This preliminary streak removal ensures that subsequent document range detection is not adversely affected by streak interference, maintaining both processing simplicity and detection accuracy.
Data Source
AI summary
An image processing apparatus includes: an input device that inputs a document image; and a controller, wherein the controller removes a streak from the image by a first method, removes a streak from the image by a second method that differs from the first method, and executes crop processing on the image from which the streaks have been removed.