Identifiable Decomposition of Submovements in Human Hand Trajectories

2026-09-30

AI Summary

This study addresses the poor scalability and lack of identifiability criteria in existing methods for decomposing human hand movements into sub-actions. We propose Sub-ID, a method that establishes the first primitive-pair-based identifiability theory, explicitly detecting indistinguishable overlaps using spatiotemporal kernel correlation as a criterion. By optimizing low-correlation solutions via adaptive ridge regression, it achieves effective decomposition of long-duration three-dimensional movements. Experiments demonstrate that Sub-ID accurately recovers ground-truth distributions on synthetic data and successfully extracts sub-action primitives from real-world 3D long-duration hand trajectories. This approach overcomes the processing limitations of conventional methods, providing physiologically grounded support for motor control research.

Abstract

Voluntary movements have long been hypothesised to be comprised of discrete primitives called submovements, as a descriptive model of human motor behaviour. However, existing methods scale poorly, and no principled method exists to determine whether a decomposition is informative. We propose a spatiotemporal kernel correlation between primitive pairs as an identifiability criterion. Submovement-Identifiable Decomposition (Sub-ID) embeds this criterion in its adaptive-ridge regularisation, biasing the optimiser toward low-correlation solutions. Identifiability is lost when primitives become collinear and recovered when they diverge spatially. On synthetic data, Sub-ID recovers ground-truth parameter distributions where existing methods fail; furthermore, when primitives overlap too heavily to be distinguished, the method explicitly detects this ambiguity rather than outputting misleading results. Sub-ID extracts submovements from real three-dimensional, long-horizon movements, a regime no prior method addresses. This method has the potential to identify physiologically grounded primitives for motor control research and imitation learning.

Problem

• submovement decomposition
• identifiability
• human hand trajectories
• motor primitives
• scalability

Innovation

• Submovement decomposition
• Identifiability criterion
• Spatiotemporal kernel correlation
• Adaptive-ridge regularisation
• Ambiguity detection