Document Image Orientation Using Reference Feature Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing systems inaccurately orient captured images of documents, leading to unnecessary resource consumption and user inconvenience due to the need for additional image captures to correct orientation.
Innovation Solution
An image processing device identifies a reference feature in the image, such as a face or machine-readable code, using a machine learning model to determine a rotational angle, and rotates the image accordingly to achieve correct orientation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image processing systems use conventional orientation methods, then processing is simpler, but orientation accuracy deteriorates leading to incorrect image display
Solution Approach 1:
The patent introduces a machine learning model as an intermediary component between the image input and orientation determination. This model acts as a mediator that automatically identifies reference features and calculates rotational angles, resolving the contradiction by providing high orientation accuracy through sophisticated processing while encapsulating the complexity within a dedicated module rather than distributing it throughout the entire system.
Solution Approach 2:
The system employs self-service by using the image content itself (reference features like faces or codes) to automatically determine orientation. The machine learning model autonomously identifies features and calculates rotation angles without requiring external intervention or complex manual processing rules, achieving high accuracy while keeping the overall system architecture relatively simple.
2Measurement precision
If additional image captures are taken to correct orientation, then orientation accuracy improves, but resource consumption increases
Solution Approach 1:
The system performs preliminary action by calculating the correct rotation angle from the first captured image using machine learning. Instead of requiring additional captures to ensure correct orientation, the system proactively determines the rotational angle and applies correction to the original image, thereby achieving high orientation accuracy while avoiding the energy consumption associated with multiple image captures.
3Reliability
If additional image captures are taken to ensure correct orientation, then orientation reliability improves, but processing time increases
Solution Approach 1:
The patent replaces the mechanical approach of physically recapturing images with a computational approach using machine learning. The machine learning model substitutes for the mechanical re-capture process by automatically analyzing the first image, determining the correct orientation through feature recognition, and applying rotational correction, thereby achieving high orientation reliability without the time loss associated with additional captures.
4Measurement precision
If conventional orientation methods are used, then processing speed is faster, but orientation accuracy deteriorates
Solution Approach 1:
The system applies parameter changes by transitioning from conventional orientation parameters (based on simple geometric or metadata cues) to machine learning-derived parameters (based on feature recognition and rotational angle calculation). This change in parameters enables high orientation accuracy while the machine learning model processes the information efficiently in a single pass, maintaining processing speed despite the increased sophistication of the orientation determination method.
Data Source
AI summary
In some implementations, a device may receive an image of a document, the image depicting a reference feature associated with the document, the reference feature including at least one of: a face of a person, a machine-readable code, or a text field. The device may identify a rotational angle of the reference feature as depicted in the image based on comparing the reference feature as depicted in the image to one or more orientation parameters of the reference feature associated with a display orientation associated with the document. The device may rotate the image of the document by an angle to obtain an orientated image of the document, the angle being based on the rotational angle of the reference feature as depicted in the image. The device may provide the orientated image of the document for display.