Object Detection via Internal Features for Mobile Document Capture
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Mobile devices face challenges in processing digital images of documents due to limited processing power, image resolution, and inherent distortions introduced by camera optics, making it difficult to detect objects accurately, especially when edges are missing or obscured, and reconstructing objects in three-dimensional space.
Innovation Solution
A computer program product that detects objects in digital images by identifying internal features, projecting regions of interest, and building an extraction model to extract content, using techniques such as feature vector analysis and homography transforms to correct for distortions and reconstruct objects in a three-dimensional coordinate space.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional scanner-based processing algorithms are used on mobile devices, then document capture and processing can be performed, but processing power requirements and computational cost become prohibitively expensive
Solution Approach 1:
The patent changes the fundamental parameters of the image processing approach by transitioning from high-computational algorithms to optimized algorithms that work efficiently with mobile device hardware capabilities, including lower resolution inputs and limited processing power while maintaining acceptable processing quality
Solution Approach 2:
The patent employs lightweight processing models and simplified algorithms that can be executed quickly on mobile devices with limited resources, sacrificing some of the heavy computational overhead of scanner-based systems in exchange for practical mobile deployment
2Power
If images are captured at lower resolutions on mobile devices, then processing becomes more feasible, but conventional scanner-based processing algorithms perform poorly
Solution Approach 1:
The patent modifies processing parameters and algorithms to be optimized for lower resolution inputs, adjusting feature detection thresholds, scaling factors, and processing sensitivity to maintain adequate detection accuracy despite reduced image quality
Solution Approach 2:
The patent applies selective processing strategies that focus computational resources on critical detection tasks rather than attempting full conventional processing, achieving sufficient accuracy for mobile applications without requiring full scanner-level precision
3Measurement precision
If edge-based detection methods are used, then object boundaries can be identified, but detection fails when edges are missing or obscured
Solution Approach 1:
The patent segments the object detection process into multiple independent feature detection stages, including edge detection, corner detection, and internal feature identification, allowing the system to rely on alternative features when edges are unavailable
Solution Approach 2:
The patent introduces intermediate feature detection mechanisms that bridge between direct edge detection and final object identification, using corner points, internal patterns, and contextual information as mediators when primary edge features are missing or obscured
4Adaptability or versatility
If camera optics are used for capture, then mobile document capture is enabled, but inherent distortions are introduced that complicate processing
Solution Approach 1:
The patent applies preliminary distortion correction and calibration procedures before main processing operations, pre-compensating for known camera optical distortions through calibration patterns and transform corrections
Solution Approach 2:
The patent replaces mechanical precision requirements with computational correction methods, using software-based distortion compensation and perspective transformation to correct optical imperfections without requiring mechanically precise capture conditions
Data Source
AI summary
A method of detecting an object depicted in a digital image includes: detecting a plurality of identifying features of the object, wherein the plurality of identifying features are located internally with respect to the object; projecting a location of region(s) of interest of the object based on the plurality of identifying features, where each region of interest depicts content; building and/or selecting an extraction model configured to extract the content based at least in part on: the location of the region(s) of interest, the of identifying feature(s), or both; and extracting the some or all of the content from the digital image using the extraction model. Corresponding system and computer program product embodiments are disclosed. The inventive concepts enable reliable extraction of data from digital images where portions of an object are obscured/missing, and/or depicted on a complex background.