Machine-Learning Check Imaging with Video-Based Authenticity Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing remote deposit capture systems face challenges in verifying the authenticity of checks and users, handling diverse environmental and hardware conditions, and addressing document security features and digital manipulation, which complicates the validation process.
Innovation Solution
A computing system that utilizes a machine-learning architecture to process image data from client devices, detecting documents like checks, extracting information, and validating authenticity through object recognition engines, bounding boxes, and dynamic alignment indicators, while generating risk scores and prompts for additional verification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If remote deposit capture is implemented to eliminate travel and staffing requirements, then convenience and accessibility are improved, but the ability to verify user identity and check authenticity deteriorates
Solution Approach 1:
The system performs preliminary actions by capturing a video feed of the check and environment before the actual deposit processing occurs. The video feed is analyzed in real-time to detect document authenticity, extract information, and verify security features before the deposit is finalized, ensuring verification happens proactively rather than reactively.
Solution Approach 2:
The patent introduces an intermediary machine learning processing system that acts as a mediator between the user's remote deposit submission and the institution's verification processes. The ML architecture including object recognition engines and anomaly detection engines serves as an intermediary layer that automatically verifies authenticity without requiring physical presence or manual inspection.
2Measurement precision
If detailed instructions are provided to users for remote deposit capture, then verification accuracy may improve, but user convenience deteriorates
Solution Approach 1:
The system enables self-service by allowing users to simply capture a video feed of their check using their mobile device camera without following complex instructions. The machine learning system automatically performs document detection, information extraction, and authenticity verification, eliminating the need for users to manually complete forms or follow detailed procedural guidance.
Solution Approach 2:
The patent replaces manual verification processes with automated machine learning systems. Instead of requiring users to follow mechanical instructions for proper document placement and lighting, the system uses computer vision and ML algorithms to automatically analyze the video feed, detect document orientation, assess image quality, and verify authenticity metrics.
3Reliability
If multiple validation checks are performed on document security features, then authenticity verification improves, but system complexity increases
Solution Approach 1:
The patent implements a universal machine learning architecture that performs multiple validation functions through a single integrated system. The same ML model processes document detection, security feature verification, information extraction, and anomaly detection simultaneously, eliminating the need for separate dedicated systems for each validation task and reducing overall system complexity.
Solution Approach 2:
The system changes parameters by using machine learning models that can dynamically adjust verification thresholds and analysis depth based on the specific document type and detected risk levels. The anomaly detection engine modifies validation parameters in real-time based on confidence scores and detected anomalies, allowing flexible verification without fixed complex procedural requirements.
4Adaptability or versatility
If electronic facsimiles of documents are accepted for remote deposit, then accessibility improves, but difficulty in inspecting document authenticity increases
Solution Approach 1:
The patent replaces manual document inspection with automated machine learning-based electronic analysis. The system uses computer vision algorithms to detect document features, verify security elements, and assess authenticity from electronic facsimiles and video feeds, eliminating the need for physical inspection while maintaining or improving detection accuracy.
Solution Approach 2:
The machine learning system serves as an intermediary between the electronic facsimile and the verification process. Rather than requiring direct human inspection of digital images, the ML architecture acts as a mediator that automatically analyzes the electronic document, extracts features, compares them against authenticity criteria, and provides verification results.
Data Source
AI summary
Presented herein are systems and methods for the employment of machine learning models for image processing. A method may include a capture of a video feed including image data of a document at a client device. The client device can provide the video feed to another computing device. The method can include, by the client device or the other computing device object recognition for recognizing a type of document and capturing an image exceeding a quality threshold of the document amongst the frames within the video feed. The method may further include the execution of other image processing operations on the image data to improve the quality of the image or features extracted therefrom. The method may further include anti-fraud detection or scoring operations to determine an amount of risk associated with the image data.