Answer Document Image Verification System
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
There is a need for accurate verification of test and survey results from standardized forms to ensure integrity and prevent inaccuracies, particularly in cases where bubble markings on answer sheets are ambiguous or require human interpretation.
Innovation Solution
A method and apparatus that utilize a computer system to automatically identify and sort responses from scanned images, with data forwarded to multiple editors for review, ensuring double-checking by different editors with distinct access levels, and including features like time and date stamping for accountability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automated scanning and identification of responses is used, then processing speed and productivity are improved, but accuracy and reliability of ambiguous responses deteriorate
Solution Approach 1:
The verification process is segmented into multiple independent review stages. Responses are divided into different categories (automatically identified vs. manually reviewed) and processed through different editorial layers, allowing automated high-speed processing of clear responses while dedicating human review capacity to ambiguous cases.
Solution Approach 2:
A computer system acts as an intermediary between automated scanning and human editors. It performs preliminary processing, identifies ambiguous responses, and routes them appropriately, bridging the gap between automated efficiency and human judgment accuracy.
2Reliability
If multiple editors review responses, then accuracy and reliability are improved, but device complexity and processing time increase
Solution Approach 1:
Responses are pre-sorted and pre-categorized by the computer system before reaching editors. Ambiguous responses are automatically identified and flagged in advance, so editors only need to review specific cases rather than processing all responses from scratch, reducing redundant complexity.
Solution Approach 2:
Different levels of review intensity are applied to different responses based on their characteristics. Clear responses receive minimal or no human review, while ambiguous responses receive intensified multi-editor review, optimizing resource allocation and reducing overall system complexity.
3Reliability
If multiple editors with distinct access levels are used, then integrity and security are improved, but device complexity and operational difficulty increase
Solution Approach 1:
Editorial access rights are segmented into different levels (first editor, second editor, quality control editor). Each level has specific, well-defined permissions and responsibilities, making the complex access control system manageable through clear segmentation of duties.
Solution Approach 2:
The system incorporates feedback loops where quality control editors review work by other editors, and discrepancies trigger additional review cycles. This feedback mechanism maintains integrity while providing structured guidance that simplifies operational complexity.
4Productivity
If automated sorting and forwarding is implemented, then productivity is improved, but loss of information or contextual nuance may increase
Solution Approach 1:
The computer system serves as an intermediary that preserves and transmits contextual information through the review process. It maintains response images, scan data, and editorial decisions in a structured format, ensuring that contextual nuance is retained and accessible throughout the automated workflow.
Data Source
AI summary
Disclosed herein, among other things, are method and apparatus for verifying answer document images. According to one embodiment of a method, data collected from a scanned image of an answer document is received, the data including an indication of whether a bubble on the answer document is marked. Responses are automatically identified to have a first review, including sorting the collected data using one or more predetermined criteria. The responses are automatically forwarded to a first editor to conduct the first review, the first review including a data entry. If the data entry of the first editor varies with the indication for a particular response, the particular response is automatically forwarded to a second editor to conduct a second review, where the first editor is different from the second editor.