Adaptive Latent Trajectory Anchoring for Action Segmentation Dataset Condensation

2026-07-10

AI Summary

This work addresses the challenges of high computational cost, over-smoothed reconstructions, and rigid temporal constraints in action segmentation dataset compression by proposing a deterministic latent mapping approach based on the Denoising Diffusion Implicit Model (DDIM). The method models action segments as continuous latent trajectories within a noise manifold and introduces an adaptive latent trajectory anchoring mechanism that dynamically allocates sparse anchor points according to reconstruction difficulty. This strategy overcomes the efficiency and expressiveness limitations of conventional optimization-based inversion techniques. Evaluated across multiple benchmarks, the proposed approach significantly outperforms existing methods, achieving segmentation performance comparable to that obtained with original uncompressed data on the Breakfast dataset using only a 2.4% compression rate.

Abstract

Dataset condensation for action segmentation synthesizes compact, informative representations of long, untrimmed video datasets. The existing approach relies on Variational Autoencoders and an iterative latent optimization; it is computationally expensive and suffers from over-smoothed reconstructions and rigid temporal constraints. This paper proposes to shift the condensation paradigm from optimization-based inversion to deterministic latent mapping. By leveraging Denoising Diffusion Implicit Models, we represent action segments as continuous trajectories anchored by sparse latent points in the noise manifold. To maximize representational efficiency, we introduce an adaptive allocation mechanism that dynamically redistributes the anchoring budget based on segment-wise reconstruction difficulty. Extensive experiments demonstrate that our framework significantly outperforms state-of-the-art methods in segmentation performance across common datasets. Notably, our approach achieves performance parity with real data training while maintaining a condensation ratio of 2.4\% on Breakfast dataset.

Problem

• dataset condensation
• action segmentation
• over-smoothed reconstructions
• temporal constraints
• computational expense

Innovation

• Dataset Condensation
• Action Segmentation
• Denoising Diffusion Implicit Models
• Latent Trajectory Anchoring
• Adaptive Allocation