Automated Occlusion Detection Using Gradient and Saliency Masks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current occlusion detection in images relies heavily on human judgment, leading to errors and inconsistencies, as it is challenging to discern between acceptable and unacceptable occlusions, resulting in inconsistent determinations.
Innovation Solution
The process involves generating and applying gradient, blur, and saliency masks to an image to identify regions that should not be occluded, using a computer system to detect objects within these regions and adjust the image if occlusions occur, thereby automating the occlusion detection and adjustment process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If human judgment is used to determine acceptable occlusions, then flexibility in decision-making is maintained, but consistency and accuracy of determination deteriorate
Solution Approach 1:
The patent replaces the mechanical system of human visual inspection and judgment with an automated computer-based image processing system. The system uses gradient masks, blur masks, and saliency maps to objectively analyze images and determine occlusion acceptance, eliminating human subjectivity while maintaining consistent decision-making criteria across all images.
Solution Approach 2:
The system enables images to evaluate themselves for occlusion acceptance. By automatically generating masks and analyzing image regions, the system allows each image to self-determine whether it meets occlusion criteria without requiring external human review, thereby achieving both consistency and adaptability.
2Productivity
If automated detection is implemented, then consistency and productivity are improved, but device complexity increases
Solution Approach 1:
The patent segments the occlusion detection task into distinct computational components: gradient mask generation, blur mask generation, saliency map creation, and final occlusion determination. Each component processes specific image features independently, which simplifies the overall system architecture while enabling parallel processing and improved productivity.
3Measurement precision
If multiple masks are applied to improve detection accuracy, then measurement precision is improved, but device complexity and processing time increase
Solution Approach 1:
The patent merges multiple mask types (gradient, blur, saliency) into a unified occlusion detection framework. Rather than treating each mask as a separate complex system, the patent integrates them into a coordinated processing pipeline where each mask contributes a specific feature analysis, collectively improving detection accuracy while managing system complexity through unified control.
Data Source
AI summary
A method includes producing a filter mask based on a blur mask and a saliency mask and identifying locations of a plurality of bounding boxes of a plurality of objects of interest in a received image. The method also includes applying the filter mask to the received image and to the locations of the plurality of bounding boxes in the received image to remove at least one object of interest from consideration. The method further includes performing a comparison of a location of a bounding box of the plurality of bounding boxes of an object of interest remaining in consideration to a predetermined safe region of the received image and generating a validation result based on the comparison.