Object Recognition Models for Subtype Distinction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing machine learning models struggle to accurately distinguish between visually similar subtypes of objects, such as different models of mobile phones, leading to false positives and inefficiencies in object recognition tasks.
Innovation Solution
A system that uses machine learning-based models to extract specific features from images, which are then input into an object recognition model to improve the accuracy of subtype identification, simplifying the training process and reducing the complexity of neural networks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing machine learning models are used for object recognition, then general object classification is achieved, but accuracy in distinguishing visually similar subtypes deteriorates
Solution Approach 1:
The patent segments the object recognition task into two distinct stages: first, a general object classifier identifies the type of object (e.g., mobile phone); second, a specialized subtype classifier distinguishes between visually similar subtypes (e.g., different iPhone models). This segmentation allows each classifier to focus on specific features, improving overall accuracy for subtype identification while maintaining general object classification capabilities.
2Measurement precision
If machine learning models attempt to distinguish between visually similar subtypes, then subtype identification capability is improved, but false positives increase
Solution Approach 1:
The patent applies preliminary action by first using a general object classifier to confirm the object type before proceeding to subtype classification. This preliminary step filters out non-relevant cases and ensures that the subtype classifier only processes images of the correct object type, thereby reducing false positives while maintaining high subtype identification accuracy.
3Measurement precision
If complex neural networks are used to improve subtype distinction, then recognition accuracy is improved, but training complexity increases
Solution Approach 1:
The patent divides the complex neural network into two separate, simpler networks: a general object classifier and a specialized subtype classifier. Each network is trained on specific datasets and features relevant to its function, reducing the overall training complexity compared to a single monolithic network while achieving superior subtype recognition accuracy through coordinated operation of both classifiers.
Data Source
AI summary
Machine learning based models recognize objects in images. Specific features of the object are extracted from the image using machine learning based models. The specific features extracted from the image assist deep learning based models in identifying subtypes of a type of object. The system recognizes the objects and collections of objects and determines whether the arrangement of objects violates any predetermined policies. For example, a policy may specify relative positions of different types of objects, height above ground at which certain types of objects are placed, or an expected number of certain types of objects in a collection.