Covariance Appearance Models for Object Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for detecting and recognizing objects of interest in images, such as those used in autonomous vehicles, face challenges due to variations in target size, distance, resolution, and operational conditions, leading to high computational intensity and costs.
Innovation Solution
A method and system that detect salient features in images, segment them into regions of interest, and generate covariance appearance models for comparison with stored models to determine if the features correspond to known objects, using a combination of salient feature detection, image segmentation, and covariance appearance modeling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional automated recognition systems are used to detect objects of interest in images, then detection accuracy can be maintained, but computational intensity and processing time increase significantly
Solution Approach 1:
The image is segmented into multiple regions of interest based on detected salient features. Each region is processed independently to generate local covariance appearance models, which reduces the overall computational burden compared to processing the entire image at once while maintaining detection accuracy through localized analysis.
Solution Approach 2:
The patent extracts and compares only the most discriminative features (covariance appearance models) from each region of interest rather than analyzing all image data. This selective extraction of key features reduces computational intensity while preserving the essential information needed for accurate detection.
2Reliability
If comprehensive feature analysis is performed to handle size variety and operational condition variations, then detection reliability improves, but computational cost and time increase
Solution Approach 1:
The patent applies different processing strategies to different regions of interest based on their specific characteristics. Each region's covariance appearance model is generated and compared according to its local features, allowing the system to handle size variety and operational condition variations effectively while optimizing computational resources for each specific region rather than applying uniform heavy processing everywhere.
Data Source
AI summary
Methods and apparatus are provided for recognizing particular objects of interest in a captured image. One or more salient features that are correlative to an object of interest are detected within a captured image. The captured image is segmented into one or more regions of interest that include a detected salient feature. A covariance appearance model is generated for each of the one or more regions of interest, and first and second comparisons are conducted. The first comparisons comprise comparing each of the generated covariance appearance models to a plurality of stored covariance appearance models, and the second comparisons comprise comparing each of the generated covariance appearance models to each of the other generated covariance appearance model. Based on the first and second comparisons, a determination is made as to whether each of the one or more detected salient features is a particular object of interest.