Incremental Feature Extraction for Object Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current object recognition systems in images face high computational complexity due to the need to extract feature descriptors from multiple octaves, which can be resource-intensive, especially on mobile devices, and often result in false positives when recognizing multiple objects in varying imaging conditions.
Innovation Solution
The system employs incremental feature descriptor extraction, starting with a selected octave and progressing only if the confidence value does not exceed a threshold, using sensors like GPS and depth estimation units to determine the starting octave and scale, and exploiting interrelationships between keypoints to reduce computational load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If feature descriptors are extracted from multiple octaves to improve object recognition accuracy, then recognition reliability is improved, but computational complexity increases
Solution Approach 1:
The patent segments the feature extraction process into discrete octaves, where each octave represents a specific scale level. The system extracts features from multiple octaves but processes them in a structured, hierarchical manner, allowing selective processing based on confidence thresholds rather than uniformly processing all octaves at full detail.
Solution Approach 2:
The patent implements partial action by using confidence thresholds to determine when to stop extracting features from additional octaves. If the confidence value exceeds the threshold after processing a certain octave, the system terminates further extraction, performing only the necessary partial work rather than exhaustively processing all possible octaves.
2Measurement precision
If feature descriptors are extracted from all octaves to ensure complete object identification, then measurement precision is improved, but loss of time increases
Solution Approach 1:
The system performs partial extraction by monitoring confidence values at each octave level and terminating the extraction process early when the confidence threshold is met, thus avoiding unnecessary processing time while maintaining sufficient identification precision for the given application.
Solution Approach 2:
The patent incorporates feedback mechanisms where confidence values computed at each octave level are used to dynamically control whether to proceed to the next octave. This feedback loop allows the system to adapt the processing depth based on the actual recognition progress, reducing time loss when high precision is already achieved.
3Productivity
If incremental feature extraction is used to reduce computational complexity, then productivity is improved, but measurement precision may deteriorate
Solution Approach 1:
The incremental extraction process performs partial feature extraction from selected octaves based on confidence thresholds, achieving sufficient precision for many practical applications without the excessive computational cost of complete multi-octave processing, thus improving productivity while maintaining acceptable precision.
Solution Approach 2:
The patent changes the parameter of octave selection by using confidence thresholds to dynamically determine which octaves require full feature extraction. This parameter adjustment allows the system to optimize between productivity and precision by adapting the extraction depth to the specific recognition scenario.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
In one example, an apparatus includes a processor configured to extract a first set of one or more keypoints from a first set of blurred images of a first octave of a received image, calculate a first set of one or more descriptors for the first set of keypoints, receive a confidence value for a result produced by querying a feature descriptor database with the first set of descriptors, wherein the result comprises information describing an identity of an object in the received image, and extract a second set of one or more keypoints from a second set of blurred images of a second octave of the received image when the confidence value does not exceed a confidence threshold. In this manner, the processor may perform incremental feature descriptor extraction, which may improve computational efficiency of object recognition in digital images.