Object Detection Scan Planning With Rule-Based Session Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems require manual determination of the minimum number of scanning sessions for creating object detection scan plans, which is inefficient and costly, especially in augmented reality applications where real-world objects need to be scanned to create digital models for interaction.
Innovation Solution
A computing system that generates an optimized scanning list by applying rules to object combinations, ensuring all objects and their states are scanned efficiently, using fractional factorial design and rule-based enumeration to minimize the number of scanning sessions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If manual determination is used to create scan plans, then flexibility and adaptability are maintained, but time consumption and cost increase significantly
Solution Approach 1:
The system performs self-service by automatically generating scan plans without requiring manual intervention. The computing system autonomously enumerates object combinations, applies rules to identify known and unscannable objects, and creates optimized scanning lists, eliminating the need for manual scan plan creation while reducing time consumption
Solution Approach 2:
The patent replaces the mechanical manual process with an automated computational system. The computing system uses algorithms to enumerate object combinations, apply detection rules, and generate scan plans, substituting the manual mechanical process of creating scan plans with an automated electronic system
2Reliability
If all possible object combinations are scanned, then complete object detection is achieved, but the number of scanning sessions becomes excessively large
Solution Approach 1:
The system extracts and removes unnecessary scanning sessions by identifying and excluding unscannable objects from the scan plan. The computing system applies rules to determine which objects can be scanned and which cannot, extracting only the essential scanning sessions needed for complete object detection while eliminating redundant ones
Solution Approach 2:
The patent changes the parameter of scan plan generation from exhaustive enumeration to optimized selection. The computing system uses rule-based filtering and object combination analysis to transform the complete set of possible scans into an optimized subset that maintains detection completeness while reducing the number of scanning sessions
3Device complexity
If scan plans are created without rules, then simplicity is maintained, but detection accuracy for known and unscannable objects decreases
Solution Approach 1:
The system performs preliminary action by pre-defining and applying rules before scan plan generation. The computing system establishes detection rules in advance that identify known objects and determine scannability, ensuring accurate object state detection is built into the scan plan creation process itself rather than added as a subsequent step
Data Source
AI summary
A system for generating a scanning list of objects to be scanned for creating an image data set for use in object detection training. A list of the objects to be scanned is received. An initial set of object combinations of the objects to be scanned is enumerated. A modified set of object combinations is generated by applying rules to each object combination in the initial set of object combinations, the rules including identifying known objects based on inclusion of other objects in the combination, and the rules including identifying unscannable objects based on inclusion of other objects in the combination. The scanning list is generated by selecting object combinations from the modified set of object combinations until all objects to be scanned appear at least once in the scanning list.