Image Feature Trackability Ranking for Video Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image-based object tracking technologies face challenges in efficiently detecting and ranking image features for tracking, leading to computational bottlenecks and reduced tracking fidelity, especially at high frame rates, as existing methods do not adequately analyze feature properties for trackability and ranking.
Innovation Solution
A method is introduced to determine the trackability of image features by establishing a tracking region within a training video sequence using feature detection algorithms, compiling trajectories, assigning saliency metrics, and employing a trackability ranking algorithm to rank features based on their trackability, which can be applied to query video sequences to select the most relevant features for tracking.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If feature detection algorithms are applied to all image features in video streams, then tracking completeness is improved, but computational processing time increases significantly
Solution Approach 1:
The patent segments the set of image features into two distinct groups: trackable features and non-trackable features. This segmentation is achieved by analyzing feature properties (such as motion characteristics, texture, and stability) and categorizing features based on their suitability for tracking. By processing only the segmented trackable features through the tracking algorithm, the system maintains tracking completeness for relevant features while significantly reducing computational processing time compared to processing all features.
Solution Approach 2:
The patent applies partial action by selectively applying feature detection and tracking algorithms only to a subset of features that exhibit trackable properties, rather than processing all features equally. The system identifies and processes approximately 10-20% of features that meet trackability criteria, achieving sufficient tracking performance for the intended application while reducing computational load by 80-90% compared to exhaustive processing of all features.
2Speed
If frame rate or frame size is increased, then video quality and temporal resolution are improved, but feature detection and tracking become computationally bound
Solution Approach 1:
The patent extracts and removes computationally intensive operations from the processing pipeline by pre-filtering features based on their trackability properties. The system extracts only the essential trackable features before applying the full tracking algorithm, effectively taking out the burden of processing non-trackable features. This extraction approach allows the system to handle higher frame rates and larger frame sizes without becoming computationally bound, as the reduced feature set requires fewer computational resources.
Solution Approach 2:
The patent performs preliminary analysis of feature properties before the main tracking process. By evaluating features in advance to determine their trackability (assessing motion characteristics, texture stability, and other properties), the system prepares a pre-screened set of candidate features. This preliminary action enables the main tracking algorithm to operate more efficiently at higher frame rates, as the preparatory filtering reduces the computational burden that would otherwise scale linearly with frame rate and frame size.
3Measurement precision
If feature detection is performed on all features, then tracking accuracy is improved, but computational overhead increases
Solution Approach 1:
The patent applies local quality by assigning different processing treatments to different features based on their individual properties. Trackable features receive full tracking processing to ensure high accuracy, while non-trackable features are either processed with simplified algorithms or excluded entirely. This localized differentiation allows the system to maintain high tracking accuracy for features that can support it, while reducing computational overhead by applying lighter processing to features that would not benefit from intensive tracking.
Data Source
AI summary
Image feature trackability ranking systems and methods are disclosed. A method of establishing a trackability ranking order from tracked image features within a training video sequence at a tracking analysis device includes establishing a tracking region within the training video sequence using a feature detection algorithm. Trajectories of tracked image features within the tracking region are compiled using a feature tracking algorithm. Saliency metrics are assigned to each one of the trajectories of tracked image features based on one or more feature property measurements within the tracking region, and a trackability ranking algorithm that is a function of the saliency metrics and a defined feature trajectory ranking associated with the training video sequence is determined, the trackability ranking algorithm being usable for ranking, based on trackability, tracked image features within another video sequence.