Barcode Recognition in Blurred Mobile Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Noise and blur significantly degrade image quality, particularly in images captured by mobile devices, where limited low-light capabilities and high noise characteristics, along with fixed focus issues, lead to severe blur and enhanced noise during de-blurring processes, making it difficult to accurately decipher de-blurred images.
Innovation Solution
A method involving pre-processing, connected components analysis, and probabilistic cost functions to identify and refine regions of interest, followed by generating idealized image intensity profiles and comparing them to blurred images to select a non-blurred version without increasing noise or artifacts, using techniques like image gradient magnitude filters and genetic algorithms to search for matching barcodes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional de-blur filters are applied to enhance image clarity, then blur is reduced, but noise and artifacts are enhanced and introduced
Solution Approach 1:
The patent segments the image processing task by first identifying the region of interest (ROI) containing the barcode, then applying de-blur filtering only to that specific region rather than the entire image. This selective approach minimizes noise enhancement in non-ROI areas while still improving barcode readability.
Solution Approach 2:
The patent applies different processing qualities to different regions: the ROI receives aggressive de-blur filtering to restore barcode details, while surrounding areas receive minimal or no filtering to preserve original quality and avoid artifact introduction. This local quality differentiation resolves the contradiction between clarity improvement and noise reduction.
2Manufacturing precision
If de-blurring level is increased to remove blur, then image sharpness improves, but artifacts such as ringing form around noise pixels
Solution Approach 1:
By segmenting the image into ROI and non-ROI portions, the patent enables high de-blurring levels in the ROI without affecting other areas. The barcode region receives maximum sharpness enhancement while artifact-prone areas remain untouched, eliminating the trade-off between sharpness and artifacts.
Solution Approach 2:
The patent introduces an intermediary step of ROI identification and masking before de-blurring. This intermediary process acts as a mediator that protects non-ROI areas from artifact generation while allowing aggressive processing in the ROI, thus resolving the contradiction between sharpness enhancement and artifact prevention.
3Manufacturing precision
If de-blurring is applied to mobile device images, then image quality improves, but noise characteristics are enhanced due to limited low light capabilities
Solution Approach 1:
The patent segments the mobile device image into ROI and background regions, applying de-blurring only to the ROI. This selective processing improves barcode quality while avoiding noise enhancement in the larger background area, effectively resolving the contradiction between quality improvement and noise control in mobile imaging.
4Loss of information
If fixed focus lens is used to capture close details, then subject detail is captured, but severe blur occurs when focal distance is violated
Solution Approach 1:
The patent performs preliminary de-blurring processing on the captured image after the fact, rather than relying on perfect focus during capture. This preliminary action restores details that were lost due to focal distance violations, allowing the system to recover subject detail even when the fixed focus lens produced severe blur.
Data Source
AI summary
Methods, systems, and apparatus, including computer program products, for identifying regions of interest in an image and identifying a barcode in a degraded image are provided. A region of interest is identified by pre-processing an image, generating a binary image based on a metric calculated on the pre-processed image, and analyzing regions of the image identified using connected components and other analysis. A barcode is identified by searching a population of barcodes, degrading ideal image intensity profiles of candidate barcodes, and comparing the degraded ideal image intensity profiles to an image intensity profile of the degraded image.