Pose and Sub-Pose Clustering for Gait Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems for identifying individuals based on gait patterns struggle with accuracy when individuals follow arbitrary walking paths or exhibit noise in skeleton data signals, leading to erroneous identification results, especially in unconstrained environments.
Innovation Solution
The method employs pose and sub-pose clustering by correlating gait features over multiple gait cycles with static and dynamic feature vectors, removing noisy frames, and creating clusters to segregate characteristic poses and sub-poses, resulting in a gait-pose feature data set that allows for unique identification regardless of activity patterns or paths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional gait-based identification systems are used, then identification can be performed non-intrusively from a distance, but identification accuracy deteriorates when individuals follow arbitrary walking paths or when noise is present in skeleton data signals
Solution Approach 1:
The patent segments the continuous gait signal into discrete pose clusters and sub-pose clusters. By dividing the gait cycle into clustered pose segments, the system can identify individuals based on characteristic pose patterns rather than continuous arbitrary motion, thereby maintaining accuracy despite variations in walking paths and noise in the skeleton data signals.
2Adaptability or versatility
If gait features are extracted for arbitrary activity patterns, then the system becomes more versatile, but measurement precision deteriorates due to noise in skeleton data signals
Solution Approach 1:
The patent extracts and removes noisy frames from the skeleton data sequence before further processing. By selectively extracting and eliminating frames with high noise levels, the system maintains measurement precision while still being able to process arbitrary activity patterns, as the extraction process preserves genuine motion information while discarding noise-corrupted data.
Solution Approach 2:
The patent changes the parameter representation from continuous gait features to discrete pose cluster labels. By transforming the continuous gait signal into discrete cluster assignments, the system becomes more robust to noise while maintaining versatility in handling arbitrary activity patterns, as cluster-based representation is less sensitive to signal variations caused by noise.
3Reliability
If pose and sub-pose clustering is implemented, then identification accuracy is enhanced and the system becomes agnostic to arbitrary activity patterns, but device complexity increases
Solution Approach 1:
The patent performs preliminary clustering of poses and sub-poses during an offline training phase, creating lookup tables of characteristic pose patterns for each individual. This preliminary action allows the online identification phase to simply match observed poses against pre-computed clusters, significantly reducing real-time processing complexity while maintaining high identification accuracy.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The subject matter discloses systems and methods for identification of individuals. The method includes obtaining static and dynamic feature vectors for skeleton data frames of each individual performing a step activity with an arbitrary pattern and in a random path; creating, for the each individual, a first predefined number of clusters of dynamic feature vectors for the frames; creating, for the each individual, a second predefined number of sub-clusters within the each of the clusters of the dynamic feature vectors for the frames associated with the each of the clusters; and determining, for the each individual, a gait-pose feature data set based on computation of a center of the dynamic feature vectors for the frames associated with the each of the sub-clusters, and a mean of the static feature vectors for the frames associated with the each of the clusters, for identifying the individuals.