Depth-Dependent Feature Detection for Scale Distinction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current computer vision methods struggle to distinguish between features at different physical scales, particularly when the scale is influenced by the distance between the camera and the object, leading to indistinguishability between real and miniature models, and are highly dependent on accurate scale computation.
Innovation Solution
A method that detects and describes features from intensity images using depth information to differentiate between real and scale-invariant features, allowing for feature detection and description at real scales without requiring dense depth data or complex 3D mesh creation, using a simple 2D intensity image for scale spaces and depth-dependent support regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If scale-invariant feature detection is used to match features across different distances, then features can be matched from different camera distances, but similar features at different physical scales become indistinguishable
Solution Approach 1:
The patent changes the parameter used for feature detection from image-space scale to depth-dependent physical scale. By incorporating depth information, the system detects features at consistent physical scales across different distances, allowing distinction between real objects and miniatures while maintaining matching capability across distances.
Solution Approach 2:
The patent introduces depth information as an intermediary parameter that mediates between image scale and physical scale. This depth data allows the system to disambiguate whether scale differences are due to distance or physical size, resolving the contradiction between distance invariance and scale distinction.
2Measurement precision
If dense depth data and complex 3D mesh creation are used to achieve real scale feature detection, then physical scale distinction is improved, but computational complexity increases
Solution Approach 1:
The patent extracts only the necessary depth information for feature detection without requiring complete 3D reconstruction. By taking out only the essential depth data needed for scale correction, the system achieves physical scale detection without the computational burden of dense depth mapping and mesh creation.
Solution Approach 2:
Instead of creating complex 3D meshes and then detecting features, the patent inverts the approach by using available depth information to directly correct feature detection scale. This reversal eliminates unnecessary processing steps while achieving the same physical scale detection goal.
3Reliability
If standard feature descriptors are used, then rotation and scale invariance are achieved, but features at different physical scales cannot be distinguished
Solution Approach 1:
The patent modifies the feature descriptor by incorporating depth-dependent physical scale information. This parameter change allows the descriptor to maintain rotation and scale invariance for matching while simultaneously encoding physical scale distinction, resolving the information loss in standard descriptors.
Data Source
AI summary
The invention provides methods of detecting and describing features from an intensity image. In one of several aspects, the method comprises the steps of providing an intensity image captured by a capturing device, providing a method for determining a depth of at least one element in the intensity image, in a feature detection process detecting at least one feature in the intensity image, wherein the feature detection is performed by processing image intensity information of the intensity image at a scale which depends on the depth of at least one element in the intensity image, and providing a feature descriptor of the at least one detected feature. For example, the feature descriptor contains at least one first parameter based on information provided by the intensity image and at least one second parameter which is indicative of the scale.