Client-Side Check Image Validation Using Native ML APIs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing mobile operations software programs for remote check deposit face challenges such as reliance on third-party libraries causing lag and performance issues, difficulty in validating document security features and user identity, and inefficiencies in image capture and validation due to non-AI methods like rule-based OCR.
Innovation Solution
Implementing a mobile application that leverages native image processing software (e.g., Apple VisionKit®) for client-side image processing and validation, using machine-learning models to identify and validate documents like checks by capturing frames, determining dimensions, and assessing brightness and contrast.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If third-party libraries are used for image processing in mobile applications, then functionality is provided, but processing speed decreases and crashes occur
Solution Approach 1:
The patent extracts the image processing functionality from third-party libraries and implements it using native image processing software (Apple VisionKit) directly on the client device. This removes the dependency on external libraries that cause crashes and performance issues, while maintaining the required functionality through built-in device capabilities
Solution Approach 2:
The mobile application leverages the native image processing capabilities already present in the client device's operating system. By using Apple VisionKit and other built-in frameworks, the application serves itself without requiring additional third-party dependencies, thereby improving both reliability and processing speed
2Measurement precision
If rule-based OCR methods are used for document validation, then implementation is straightforward, but accuracy and efficiency are insufficient
Solution Approach 1:
The patent replaces rule-based OCR methods with machine learning models that use neural networks for document validation. This substitution transitions from deterministic rule-based processing to probabilistic AI-based processing, significantly improving accuracy in identifying document types and validating security features despite increased computational requirements
Solution Approach 2:
The system changes the fundamental parameters of document validation by transitioning from simple text recognition to multi-parameter analysis including document geometry, visual patterns, security feature detection, and contextual information. This comprehensive parameter analysis enables more accurate validation while the patent manages complexity through efficient model deployment on mobile devices
3Measurement precision
If detailed instructions are provided to users for remote deposit capture, then validation accuracy may improve, but user convenience decreases
Solution Approach 1:
The patent implements real-time feedback mechanisms where the machine learning model continuously analyzes the captured image and provides immediate guidance to the user. The system feedbacks information about proper document placement, lighting conditions, and capture quality, allowing users to self-correct without detailed instructions while maintaining high image quality standards
Solution Approach 2:
The mobile application performs automatic image quality assessment and validation without requiring user intervention or detailed instructions. The system independently evaluates captured images against quality criteria and either accepts or requests recapture, thereby maintaining convenience while ensuring accuracy through AI-driven automated quality control
4Reliability
If multiple security features are validated, then document authenticity improves, but processing time increases
Solution Approach 1:
The patent performs preliminary analysis of document security features during the image capture phase using machine learning models. By pre-validating geometric properties, visual patterns, and security feature presence before full processing, the system reduces subsequent validation time while maintaining comprehensive authenticity verification
Solution Approach 2:
The validation process is segmented into multiple independent analysis stages: geometric validation, visual pattern recognition, security feature detection, and contextual verification. Each segment processes specific aspects of document authenticity in parallel, reducing overall processing time through modularization while maintaining comprehensive validation
Data Source
AI summary
Presented herein are systems and methods for the employment of machine learning models for image processing. A mobile application for client-side image processing and validation, which interacts with and leverages native image processing software of the client device, where the image processing software and the mobile application include any number of machine-learning models for identifying a document and attributes of the document for recognition and validation. This mobile application uses the image processing software from a client operating system to control the camera. The image processing software generates various types of information about a video frame and the document, and the mobile application invokes APIs or software libraries of the image processing software to access the information and validate the frame and document.