Large-Scale Object Detector for Unknown Classes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image editing systems are limited in flexibility, accuracy, and efficiency when detecting and selecting objects in digital images, particularly failing to recognize objects from unknown classes and requiring excessive computational resources.
Innovation Solution
The implementation of a large-scale object detector system that utilizes a region proposal model, concept mask model, and auto tagging model to detect and select objects in digital images, including those from unknown classes, by generating region proposals, filtering them based on a concept mask, and tagging remaining proposals to accurately identify the query object.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional object detection systems are used, then processing speed is maintained, but detection accuracy and flexibility for unknown object classes deteriorate
Solution Approach 1:
The system segments the object detection process into three distinct modules: region proposal generation, concept mask filtering, and auto-tagging. This segmentation allows each module to specialize in specific tasks, improving overall detection accuracy for unknown objects while maintaining processing efficiency.
Solution Approach 2:
The patent introduces concept masks as an intermediary element between region proposals and final object identification. These concept masks serve as a bridge that filters and refines potential object regions before auto-tagging, enabling the system to accurately detect unknown object classes without sacrificing processing speed.
2Adaptability or versatility
If comprehensive object detection for all classes is implemented, then detection versatility improves, but computational resource consumption increases
Solution Approach 1:
The system performs preliminary filtering using concept masks before applying computationally intensive auto-tagging operations. This preliminary action reduces the number of candidate regions that require full processing, thereby detecting all object classes while conserving computational resources.
Solution Approach 2:
The patent applies partial processing to region proposals by first filtering through concept masks and then applying auto-tagging only to filtered candidates. This partial action approach ensures comprehensive detection coverage while avoiding excessive computational resource consumption on all possible regions.
3Measurement precision
If multiple processing stages are added to improve accuracy, then object detection precision improves, but system complexity increases
Solution Approach 1:
The system divides the complex object detection task into three manageable segments: region proposal, concept mask filtering, and auto-tagging. This segmentation improves identification accuracy while keeping each individual module relatively simple and well-defined.
Solution Approach 2:
The concept mask model serves multiple functions: it filters region proposals, provides semantic guidance for auto-tagging, and reduces computational load. This multi-functionality improves object identification accuracy without proportionally increasing system complexity.
Data Source
AI summary
The present disclosure relates to an object selection system that automatically detects and selects objects in a digital image utilizing a large-scale object detector. For instance, in response to receiving a request to automatically select a query object with an unknown object class in a digital image, the object selection system can utilize a large-scale object detector to detect potential objects in the image, filter out one or more potential objects, and label the remaining potential objects in the image to detect the query object. In some implementations, the large-scale object detector utilizes a region proposal model, a concept mask model, and an auto tagging model to automatically detect objects in the digital image.