Reference-Feature Document Cropping for Hard-to-Detect Edges
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing document detection techniques often fail to accurately identify the edges of documents in images, particularly when they are part of a book or booklet, or when the background has a similar color or intensity, leading to inefficient use of processing, computing, and network resources.
Innovation Solution
The use of a reference feature associated with defined location, size, and orientation parameters to detect the outline of a document in an image, allowing for accurate cropping without relying on contour-based edge detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If contour-based edge detection is used to identify document edges, then the detection process can be performed, but the detection accuracy deteriorates when documents have similar color backgrounds or are part of booklets
Solution Approach 1:
The patent introduces a reference feature as an intermediary element to facilitate document detection. Instead of directly detecting document edges through contour-based methods, the system first detects a known reference feature within the document, then uses the predefined spatial relationship between this reference feature and the document edges to calculate and identify the document boundaries. This intermediary approach resolves the contradiction by providing a reliable detection pathway that works even when direct edge detection fails due to similar colors or booklet contexts.
Solution Approach 2:
The patent replaces the mechanical contour-based edge detection system with an information-based detection system. Instead of relying on visual edge contours that may be indistinguishable from the background, the system substitutes a detection method based on identifying a reference feature and using computational geometry (predefined location parameters) to derive document edges. This substitution transforms the detection problem from one of visual edge identification to one of reference feature identification and mathematical calculation, thereby improving reliability in challenging visual conditions.
2Measurement precision
If additional image captures and processing are performed to improve document detection accuracy, then detection precision may improve, but processing time and computational resources increase
Solution Approach 1:
The patent applies preliminary action by pre-defining the location parameters that describe the spatial relationship between the reference feature and the document edges. These parameters are established beforehand based on the known structure and layout of the document type. During actual detection, the system simply needs to identify the reference feature and apply these pre-computed parameters to calculate document boundaries, eliminating the need for complex real-time processing or multiple image captures. This preliminary preparation resolves the contradiction by enabling high precision detection through simple, fast operations.
3Productivity
If the entire image including book pages or background is transmitted, then no additional processing is needed, but network resources and data transmission volume increase
Solution Approach 1:
The patent applies the extraction principle by isolating and extracting only the relevant document portion from the larger image. After detecting the document edges using the reference feature method, the system extracts just the document area and transmits only this cropped region to the server. This eliminates the need to transmit entire book pages or background areas, thereby reducing network resource consumption and data transmission volume while maintaining processing efficiency. The extraction of only necessary information resolves the contradiction between productivity and energy loss.
Data Source
AI summary
In some implementations, a device may receive the image of the document, the image of the document depicting the reference feature, and the reference feature being associated with one or more location parameters for a document type associated with the document. The device may detect a location of the reference feature as depicted in the image, the location being defined by bounds of the reference feature as depicted in the image. The device may detect a border of the document as depicted in the image based on identifying one or more edges of the document based on the bounds of the reference feature and the one or more location parameters. The device may modify the image of the document based on the border of the document to obtain a cropped image. The device may transmit, to a server device, the cropped image.