Receipt Image Correction via Symbol-Based Positioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in accurately scanning and processing receipts of varying sizes and shapes, particularly in recognizing and stitching images of receipts marked with symbols, which limits their use in accounting and traceability applications.
Innovation Solution
A system and method using a camera-equipped user device to capture images of receipts, recognize symbols, and perform error correction, allowing for the stitching of images from multiple frames to create a coherent image, even when receipts are not fixed in size or orientation, utilizing markers for edge detection and orientation, and employing a stencil for enhanced image capture.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If receipts are scanned using existing technologies, then scanning can be performed, but accuracy is insufficient for receipts of varying sizes and shapes
Solution Approach 1:
The patent changes the parameter of reference framework from fixed grid to dynamic symbol-based positioning. By detecting symbols at expected locations and using them as reference points, the system adapts to receipts of varying sizes and shapes while maintaining high scanning accuracy through parameter transformation rather than fixed geometric constraints
Solution Approach 2:
The patent uses a predefined stencil pattern that copies the expected symbol layout onto the receipt image. This virtual stencil is overlaid on the captured image to locate symbols at expected positions, enabling accurate reference point identification regardless of the actual receipt dimensions or orientation
2Stability of the object's composition
If symbols are recognized for stitching images, then image coherence can be achieved, but image errors occur due to distortion and blurring
Solution Approach 1:
The patent implements feedback by detecting actual symbol locations and comparing them with expected stencil positions. The system uses this feedback to calculate transformation parameters that correct for distortion and blurring, iteratively refining the reference framework until image coherence and accuracy are achieved
Solution Approach 2:
The patent performs preliminary error correction by detecting symbols and calculating transformation parameters before final image stitching. This preliminary action identifies and corrects distortion and blurring issues in advance, ensuring reliable reference point establishment for subsequent stitching operations
3Area of stationary object
If multiple images are stitched together, then complete receipt capture is achieved, but processing time increases
Solution Approach 1:
The patent performs preliminary symbol detection and reference point identification on individual frames before stitching. By pre-processing each frame to locate symbols and establish reference points in advance, the system reduces the computational burden during the stitching phase, achieving complete receipt capture with minimized processing time
Solution Approach 2:
The patent segments the receipt scanning process into independent frame captures with pre-identified reference points. Each frame is processed separately to detect symbols and calculate transformations, then efficiently stitched together using these pre-computed parameters, reducing overall processing time compared to processing the entire receipt as a single image
Data Source
AI summary
A camera may capture data. A processor in communication with the camera may detect a plurality of symbols in one or more frames in the data. The processor may determine an expected sequence of the plurality of symbols and an expected orientation of each of the plurality of symbols. The processor may determine a position and orientation of each of the one or more frames based on at least one of the symbols visible in the frame. The processor may correct errors in the one or more frames. The processor may arrange a plurality of frames with respect to one another based on the determined positions and orientations. The processor may stitch the plurality of arranged frames into an image.